Sie sind auf Seite 1von 159

Designing

HPE Backup Solutions eBook


(Exam HPE0-J77)
First Edition
Aleksandar Miljkovi

HPE Press
660 4th Street, #802
San Francisco, CA 94107

Designing HPE Backup Solutions eBook (Exam HPE0-J77)


Aleksandar Miljkovic

2016 Hewlett Packard Enterprise Development LP.

Published by:
Hewlett Packard Enterprise Press
660 4th Street, #802
San Francisco, CA 94107

All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or by any information storage and retrieval
system, without written permission from the publisher, except for the inclusion of brief quotations in a
review.

ISBN: 978-1-942741-32-9

WARNING AND DISCLAIMER


This book provides information about the topics covered in the Designing HPE Backup Solutions (HPE0J77) certification exam. Every effort has been made to make this book as complete and as accurate as
possible, but no warranty or fitness is implied.
The information is provided on an as is basis. The author, and Hewlett Packard Enterprise Press, shall
have neither liability nor responsibility to any person or entity with respect to any loss or damages arising
from the information contained in this book or from the use of the discs or programs that may accompany
it.
The opinions expressed in this book belong to the author and are not necessarily those of Hewlett Packard
Enterprise Press.

TRADEMARK ACKNOWLEDGEMENTS
All third-party trademarks contained herein are the property of their respective owner(s).

GOVERNMENT AND EDUCATION SALES


This publisher offers discounts on this book when ordered in quantity for bulk purchases, which may
include electronic versions. For more information, please contact U.S. Government and Education Sales
1-855-447-2665 or email sales@hpepressbooks.com.

Feedback Information
At HPE Press, our goal is to create in-depth reference books of the best quality and value. Each book is
crafted with care and precision, undergoing rigorous development that involves the expertise of members

from the professional technical community.


Readers feedback is a continuation of the process. If you have any comments regarding how
we could improve the quality of this book, or otherwise alter it to better suit your needs, you can contact
us through email at hpepress@epac.com. Please make sure to include the book title and ISBN in your
message.
We appreciate your feedback.

Publisher: Hewlett Packard Enterprise Press
HPE Contributors: Ralph Luchs, Ian Selway, Chuck Roman
HPE Press Program Manager: Michael Bishop

About the Author


Aleksandar Miljkovi has over sixteen years of IT experience with a strong background in UNIX systems
and network security. His specialties include IT security, OpenStack based cloud computing, and HPE
Helion CloudSystem, a fully integrated, IaaS and PaaS cloud solution for the enterprise. Aleksandar is
currently involved in developing and delivering sales and technical training in areas such as
virtualization, hybrid cloud, and backup and recovery systems and solutions.

Introduction
This study guide helps you prepare for the Designing HPE Backup Solutions exam (HPE0-J77) that is a
requirement for the HPE ASE - Storage Solutions Architect V2 certification. The exam tests your ability
to define and recommend an effective Enterprise backup and recovery solution based on customer needs.
Designed to help you identify the best components, this guide describes HPE's Backup, Recovery and
Archive (BURA) solutions for data protection, and outlines recommended configurations for HPEs
StoreOnce Backup Systems.

Certification and Learning


Hewlett Packard Enterprise Partner Ready Certification and Learning provides end-to-end continuous
learning programs and professional certifications that can help you open doors and succeed in the New
Style of Business. We provide continuous learning activities and job-role based learning plans to help you
keep pace with the demands of the dynamic, fast paced IT industry; professional sales and technical
training and certifications to give you the critical skills needed to design, manage and implement the most
sought-after IT disciplines; and training to help you navigate and seize opportunities within the top IT
transformation areas that enable business advantage today.
As a Partner Ready Certification and Learning certified member, your skills, knowledge, and real-world
experience are recognized and valued in the marketplace. To continue your professional and career
growth, you have access to our large HPE community of world-class IT professionals, trend-makers and
decision-makers. Share ideas, best practices, business insights, and challenges as you gain professional
connections globally.
To learn more about HPE Partner Ready Certification and Learning certifications and continuous learning
programs, please visit
http://certification-learning.hpe.com

Audience
This book is designed for presales and solutions architects involved in supporting the sale of enterprise
backup solutions. It is assumed that you have a minimum of one year of experience in storage technologies
and have a need to identify and define the right components for a backup and recovery solution based on
customer needs.

Assumed Knowledge
To understand the concepts and strategies covered in this guide, storage professionals should have at least
six months to one year of on the job experience. The associated training course, which includes numerous
practical examples, provides a good foundation for the exam, but learners are also expected to have at
least one years experience in storage technologies.

Relevant Certifications
After you pass these exams, your achievement may be applicable toward more than one certification. To
determine which certifications can be credited with this achievement, log in to The Learning Center and
view the certifications listed on the exams More Details tab. You might be on your way to achieving
additional certifications.

Preparing for Exam HPE0-J77


This self-study guide does not guarantee that you will have all the knowledge you need to pass the exam.
It is expected that you will also draw on real-world experience designing and implementing storage
solutions. We also recommend taking the Designing HPE Backup Solutions instructor led one-day training
course (Course ID 01059534).

Recommended HPE Training


Recommended training to prepare for each exam is accessible from the exams page in The Learning
Center. See the exam attachment, Supporting courses, to view and register for the courses.

Obtain Hands-on Experience


You are not required to take the recommended, supported courses, and completion of training does not
guarantee that you will pass the exams. Hewlett Packard Enterprise strongly recommends a combination
of training, thorough review of courseware and additional study references, and sufficient on-the-job
experience prior to taking an exam.

Exam Registration
To register for an exam, go to
http://certification-learning.hpe.com/tr/certification/learn_more_about_exams.html

1 Downtime and Availability

CHAPTER OBJECTIVES
In this chapter, you will learn to:
Describe and differentiate the terms downtime and availability
Explain the business impact of system downtime
Describe and calculate the Mean Time between Failures (MTBF) and estimate its impact on the IT
environment
Describe the Mean Time to Repair (MTTR) and explain its importance in the overall recovery
List and describe technologies that address availability, such as Redundant Array of Inexpensive (or
Independent) Disks (RAID), snapshots, and replication, and contrast them with backup
Explain and use Recovery Point Objective (RPO) and Recovery Time Objective (RTO) as parameters
for backup planning

Introduction
When it comes to downtime, availability, fault tolerance, backup, recovery, and data archiving, several
topics come to mind. Almost everyone knows something about them, usually firsthand. Who has not faced
a system failure, a loss of data, or some type of a disaster that affected normal operations of a network,
the Internet, a point-of-sale system, or even your own computer?
All these terms deal with protecting business or personal assetseither systems, applications, processes,
data, or the ability to conduct business. Each varies, however, in its approach and methodology.
This chapter lays the foundation for this course and the subsequent chapters by defining and describing the
key concepts behind the quest for protecting IT resources. It begins by describing and differentiating
between downtime and availability. It then explains how a component or a system failure, which leads to
downtime, is measured and prevented. Lastly, it takes the opposite view, one of system availability, and
how it can be achieved. Within the context of both downtime and availability, it positions backup and
recovery.

Downtime
Let us begin by focusing on what prevents a system from operating normally and what the consequences
might be.

Defining downtime
Downtime is a period of time during which a system, an application, a process, a service, or data is
unavailable. During this time, a business entity cannot conduct sales operations, support a customer,
provide a service, or conduct transactions, that is, if this business entity solely relies upon what just
became unavailable.
There are many kinds of downtime, but all cause some type of disruption to normal business operations.

These events cause loss of productivity, inability to communicate with clients, or even loss of sensitive or
important information. Downtime of an IT system can cause loss of productivity of a single user, a
workgroup, or even the entire company. Whenever downtime impairs business, the outage carries serious
consequences.
To prevent or minimize the impact of downtime, you must understand what causes downtime in the first
place.

Describing causes of downtime

Figure 1-1 Causes of downtime and data loss or unavailability

Figure 1.1 lists the most typical causes of downtime and data loss or unavailability. Note that the largest
cause of downtime is due to hardware failures and then to human error. Other causes, not shown in this
chart, include the following:

Power outages
Data center cooling issues
Software bugs
Cyber attacks
Database inconsistencies or design flaws
_____________________________________ (add your own)
_____________________________________ (add your own)
_____________________________________ (add your own)

Disasters, without a doubt, affect computer operations and data unavailability. A disaster in the IT world

is any event that disrupts a companys computer operations and causes downtime. Disasters are not
controllable, but precautions against them make the recovery much faster and easier.
Disasters can be categorized according to the affected area:
Building-level incidents
Metropolitan area disasters
Regional events
Furthermore, downtime can be planned as well as unplanned.

Building-level incidents

Figure 1-2 Disasters affecting a building

Disasters affecting a building (Figure 1.2) usually impact computer operations in that building as well.
There may not be a direct damage to the IT systems, but these incidents may prevent access to them or to
the data they host, or they may interrupt operations.

Metropolitan area disasters

Figure 1-3 Metropolitan area disasters

Usually floods, fires, large chemical incidents, moderate earthquakes, severe winter storms, or blackouts
affect entire cities, impacting their infrastructure and disrupting IT systems (Figure 1.3).

Regional events

Figure 1-4 Natural disasters

Computer operations may be interrupted by natural disasters that affect an entire region within a radius of
hundreds to tens of thousands of miles/kilometers. These disasters include large floods, hurricanes,
earthquakes, political instability, and wars.

Planned vs. unplanned downtime


Planned downtime occurs when system administrators intentionally restrict or stop the system operations
to implement upgrades, updates, repairs, or other changes. During planned downtime, a particular time

period is set aside for these operations, which are carefully planned, prepared, executed, and validated.
On the contrary, unplanned downtime is when an unintentional intervention restricts or stops the system
availability.
Planned downtime means downtime as well.
Return, for a moment, to Figure 1.1, which shows the major causes of downtime and data
loss/unavailability. Are these causes intentional/planned or unintentional/unplanned? If you answered
unintentional/unplanned, you answered correctly.
Overall, the majority of system and data unavailability is because of planned downtime due to required
maintenance and upgrades. In fact, unplanned downtime accounts for only about 10% of all downtime, but
its unexpected nature means that any single downtime incident may be more damaging to the enterprise,
both physically and financially, than many occurrences of planned downtime.
Thus, understanding the cost of downtime is critical in either case.

Estimating the cost and impact of downtime


Quantifying downtime is not an easy task because the impact of downtime varies from one case to another.
Losing a second of time in an air traffic control system or in a hospital life-support environment can have
dire consequences, while losing hours in a billing system may not have a significant impact at all if these
billing transactions were queued and committed only when the system became available again.
Before you can calculate downtime, you must know its root cause. And not all root causes are strictly IT
issues. To begin with, it is important to identify and understand both internal and external downtime
threatsyou need to know what and who has the potential to take your business down. You also need to
know how.
IT-related outages, planned or unplanned, can unleash a procession of costs and consequences that are
direct and indirect, tangible and intangible, short term and long term, and immediate and far-reaching.

Tangible and direct costs


Tangible and direct costs refer to expenses that can easily be measured and documented, are incurred up
front, and are tracked in the business general ledger.
These costs can be (touched) or (felt) or (easily determined.)
Tangible and direct costs related to downtime include:

Loss of transaction revenue


Loss of wages due to employees idle time
Loss of inventory
Remedial labor costs
Marketing costs
Bank fees
Legal penalties from inability to deliver on service-level agreements (SLAs)

Intangible and indirect costs


Intangible and indirect costs refer to business impact that is more difficult to measure, is often felt or
incurred at a later date, and is not related to a physical substance or intrinsic productive value.
These costs are nonetheless real and important to a business success or failure. They can be more
important and greater than tangible costs.
Intangible and indirect costs related to downtime include the following:

Loss of business opportunities


Loss of employees and/or employee morale
Loss of goodwill in the community
Decrease in stock value
Loss of customers and/or departure of business partners
Brand damage
Shift of market share to competitors
Bad publicity and press

Outage impact to business


The cost that can be assigned to a measurable period of downtime varies widely depending upon the
nature of the business, the size of the company, and the criticality of the IT system related to the primary
revenue-generating processes. For example, a global financial services firm may lose millions of dollars
for every hour of downtime, whereas a small manufacturer using IT as an administrative tool would lose
only a margin of productivity.
According to a Gartner document titled How Much Does an Hour of Downtime Cost?, for a conventional
brick-and-mortar business, estimating the cost of an outage is relatively simple, compared to, let us say, a
global financial services firm. In either case, such estimations are never trivial.
Consider this example:
Assume that a firm conducts business in Western Europe and North America during regular business hours. This firm needs its
systems and services to be available 40 hours per week, or 2000 hours per year (accounting for holidays, vacations, and
weekends).
Therefore, the first order of approximation for the cost of an outage would be to distribute the firms revenue uniformly across
those 2000 hours. Thus, if the firms annual revenue is $100 million, an average hour would represent $50,000 of its revenue.
Consequently, one-hour outage would cost this firm $50,000.
Two immediate objections arise to this assessment and both lead to important refinements. First, revenue is almost never
distributed uniformly across all working hours. Second, most businesses experience seasonal fluctuations. Many retail
organizations make 40% of their revenue and 100% of their profits in the last eight weeks of the year. One-hour outage on
December 23 will have a much greater impact on the firms financial performance as compared to the same outage in late
June, for instance. Therefore, the cost of an outage must reflect its potential impact at a particular time in the business cycle.

Table 1.1 shows the cost of downtime, estimated in 1998 by the Gartner Group. You can see that the cost
of downtime is company-specific and related intangible costs are very difficult to estimate.

Table 1-1 The cost of downtime, estimated in 1998 by the Gartner Group
Industry

Application

Average cost (per hour of downtime)

Financial

Brokerage operations

$6,500,000

Financial

Credit card sales

$2,600,000

Media

Pay-per-view

$1,150,000

Retail

Home shopping (TV)

$113,000

Retail

Catalog sales

$90,000

Transportation

Airline reservations

$89,500

Consequences of downtime
What happens to an organization when its system goes down?
Before you can predict downtime and plan its prevention, you must have a clear understanding of what
happens when a system goes down. Specifically, what and who is impacted and how.
You should consider the following:
Processes: Vital business processes such as order management, inventories, financial reporting,
transactions, manufacturing, and human resources may be interrupted, corrupted, or even lost.
Programs: Revenue can be affected and a key employee or customer activities might be missed or
lost.
Business: If customers cannot access a website, they might purchase from someone else. You might
lose a customer now and forever.
People: Salaries might not be paid, and even lives could be lost due to downtime of life-sustaining
medical systems.
Projects: Thousands of person-hours of work can be lost and deadlines can be missed, resulting in
failure-to-perform fees and noncompliance penalties.
Operations: Those who manage daily activities of an organization may find themselves without the
data they need to make informed decisions.
The Gartner Group recommends estimating the cost of an outage to your firm by calculating lost revenue,
lost profit, and staff cost for an average hourand for a worst-case hourof downtime for each critical
business process. Not-for-profit organizations cannot calculate revenues or profits, so they should focus
on staff productivity and qualitative assessment of the outage to their user community. An outage can
weaken customer perception of the firm, harm the wider community, and derail a firms strategic
initiatives, but these impacts may be difficult to quantify, and should, in most cases, be left unquantified.
Any outage assessment based on raw, generic industry averages alone is misleading.

Predicting downtime
Can you predict downtime and events that lead to it? If you could, would it mean that you could then better
prepare for such events, or even prevent them from happening?
There are causes of downtime which cannot be predicted, such as natural disasters or fire. You just have
to determine the best prevention or recovery mechanism if they occur. On the other hand, failure rates of

computer system components can be predicted with a level of certainly. To determine and express failure
rates of computer components, you can calculate the MTBF.

Defining MTBF
MTBF is the measure of expected failure rates. To understand MTBF, it is best to start with something
elsesomething for which it is easier to develop an intuitive feel.
Let us take a look at a generalized MTBF measure for a computer component, such as a hard drive, a
memory DIMM, or a cooling fan. This component has an MTBF of 200,000 hours. Since there are
approximately 9000 hours in a year, 200,000 hours is about 21 years. In other words, if the MTBF of this
component is 200,000 hours, it is expected that the component fails every 21 years.
Now, take a sample of 10,000 units of this particular component and determine how many units fail every
day, over a test period of 12 months. You may determine that
In the first 24 hours, two components fail.
In the second 24 hours, zero components fail.
In the third 24 hours, one component fails, and so on.
You then ask
If five units fail in 12 months, how long would it take for all 10,000 units to fail at this rate?
If all units fail in regular intervals over this period, how long is this interval?
If a failure is the termination of the components ability to perform its intended function, what is then
MTBF? MTBF is an interval of time used to express the expected failure rate of a given component. It
does not indicate the expected lifetime of that component and says nothing about the failure likelihood of a
single unit.

Figure 1-5 Reliability bell curve: number of failures vs. time distribution

This reliability bell curve (Figure 1.5) is also known as a normal distribution, where the highest point of
the curve (or the top of the bell) represents the most probable event (at time ). All possible events are
then distributed around the most probable event, which creates a downward slope on each side of the

peak.
Given the reliability bell curve and a sample of components, most of them will fail around time . Some
components will fail early in the cycle, whereas others will last much longer. But all of them will fail at
some point on this curve, with a high probability that the majority of failures will be distributed according
to the bell curve.

Note
Trent Hamm has written a good explanation of the reliability bell curve in August 2014, using the
analogy of low-end and more reliable washing machines. You can find this article at:
http://www.thesimpledollar.com/the-reliability-bell-curve-what-does-more-reliability-actuallymean/.

Calculating MTBF
How does one calculate an MTBF for an environment which consists of a number of such components?
What if all these components were identical? What if they all were different?
Once again, let us use an example:
If you have 2000 identical units in your environment and each unit has the MTBF of 200,000 hours,
what is the associated total MTBF?
The formula for the total MTBF is as follows:
MTBFtotal
Therefore,
MTBFtotal

= 100 hours = ~4 days

Within your particular environment of 2,000 identical units, you can expect a failure approximately every
4 days.
It may be easy to understand the MTBF of a single unit, but in reality, systems are complex and consist of
a number of different components. Within a complex environment, you can expect a failure of any
particular component, a failure between components, or a failure of the entire system.
The measure of a failure rate for a system consisting of one to n components, which may not necessarily
be identical, is expressed as the MTBF for a complex system. Its formula is as follows:
MTBF complex system=
Even though individual MTBF rates for a single component have been improving, the entire environment
is still vulnerable to component failures due to the interdependent nature of complex and multivendor
solutions. Consider these examples:
Hardware: Storage systems may fail to provide adequate service due to MTBF problems or due to
corrupted data (resulting from viruses or data integrity issues). Computing systems may lose power

or can have various other problems, such as memory or processor failures. One corner stone for a
highly available environment is a good network. High-speed connections for transaction processing
and robust connections for client/server backup are critical.
Software: The basis of any system is a stable operating system because a failure at the OS level may
lead to vulnerability and data loss in everything that is at higher levels of the stack, such as
applications, databases, and processes.

Defining and calculating MTTR


You are looking at a component or a system failure rate in order to predict, prevent, and minimize
downtime. There is another variable which has not yet been discussed, one of recovery. As soon as a unit
fails, how long does it take to repair or replace it? To answer this question, another term is used, called
MTTR (Mean Time to Repair or Mean Time to Recovery).
MTTR represents the average time it takes to repair or replace the defective component and return the
system to its full operation. It includes the time needed to identify the failure, to diagnose it and determine
its root cause, and to rectify it.
MTTR can be a major component of downtime, especially if you are unprepared.

Planning for and preventing downtime


While planning for downtime, you have to be aware of the level of problem you are planning for. While
you cannot influence metropolitan or regional incidents because they are usually naturally occurring
disasters, you can create an appropriate disaster recovery plan and implement a disaster recovery
solution once you know what it is you are planning for (also called local problem area). Many times, you
can also put in place a solution that may reduce the impact of a disaster or even prevent it altogether.
Lastly, regardless of the cause, location, or the level of downtime, some type of administrative
intervention is usually required to quickly recover from the problem. This intervention always depends on
the exact cause of the problem and differs from one case to another. To increase your chances of success,
make sure that you have

Adequate problem detection in place


A recovery plan that is validated and tested
Protection technologies that ensure desired levels of availability
Monitoring and diagnostic tools and processes
Skilled staff educated in the adopted technologies and recovery processes
_____________________________________ (add your own)
_____________________________________ (add your own)
_____________________________________ (add your own)

Learning check questions


Reinforce your knowledge and understanding of the topics just covered by completing this learning check:

1. What measures the time needed to repair a component after it fails?


a)MTBF
b)MTTR
c)RTO
d)RPO
2. What measures the MTBFs of a complex system?
a)MTBF
b)MTTR
c)RTO
d)RPO
3. Floods, wildfires, tornados, and severe snowstorms are examples of what type of disasters?
a)Building-level incidents
b)Metropolitan area disasters
c)Regional events
d)Localized incidents
4. What is the major cause of system and data unavailability?
a)User errors
b)Metropolitan area disasters
c)Planned downtime
d)Inadequate backup strategy

Learning check answers


This section contains answers to the learning check questions.
1. What measures the time needed to repair a component after it fails?
a)MTBF
b)MTTR
c)RTO
d)RPO
2. What measures the MTBFs of a complex system?
a)MTBF
b)MTTR
c)RTO
d)RPO
3. Floods, wildfires, tornados, and severe snowstorms are examples of what type of disasters?

a)Building-level incidents
b)Metropolitan area disasters
c)Regional events
d)Localized incidents
4. What is the major cause of system and data unavailability?
a)User errors
b)Metropolitan area disasters
c)Planned downtime
d)Inadequate backup strategy

Availability
Up to now, this module covered the effects of downtime of a computer system or unavailability of data
due to a failure or a disaster. The opposite of downtime is availability (or uptime).
This section discusses the terminology and methods of expressing and calculating the availability times
and requirements.

Learner activity
Before continuing with this section, see if you can define the terms below and answer the questions.
Search the Internet if you need to.
In your own words, define these terms:
o High availability:
__________________________________________________
__________________________________________________
__________________________________________________
__________________________________________________
o Fault tolerance:
__________________________________________________
__________________________________________________
__________________________________________________
__________________________________________________
o RPO:
__________________________________________________
__________________________________________________
__________________________________________________
__________________________________________________
o RTO:
__________________________________________________

__________________________________________________
__________________________________________________
__________________________________________________
Given the following type of failure or downtime, what would be the best prevention technology in
your opinion?
o Power outage: ____
o Hardware component failure: ____
o System-wide failure: ____
o Administrative or maintenance downtime: ____
o Regional or metropolitan disaster: ____

Defining availability
Availability is a measure of the uptime of a computer system. It is a degree to which a system, a
subsystem, or a component of a system continues running and being fully operational. High availability is
something which is generally acceptable and desirable. Higher levels of availability, all the way to
continuous or mission-critical availability, are usually expensive and complex to achieve.

Categorizing availability
When a computer system plays a crucial role, availability requirements are dictated and enforced by the
law or technical regulations. Examples of such systems include the hospital life-support equipment, air
traffic systems, and nuclear power plant control systems.
The Harvard Research Group defines five Availability Environments (AEs) in terms of the impact on the
business and the end user/customer. Each successive level inherits the availability and functionality of the
previous (lower) level.
The minimum requirement for a system to be considered highly available is a backup copy of data on a
redundant disk and a log-based or journal file system for identification and recovery of incomplete ((in
flight)) transactions that have not been committed to a permanent medium. This environment corresponds
to AE-1.

Figure 1-6 AEs as defined by the Harvard Research Group

Figure 1.6 summarizes these five AEs. They are defined as follows:
AE-4 (fault tolerance)AE-4 covers business functions that demand continuous computing and
environments where any failure is transparent to the user. This means no interruption of work, no
lost transactions, no degradation in performance, and continuous 247 operation.
AE-3 (fault resilience)AE-3 covers business functions that require uninterrupted computing
services, either during essential periods or during most hours of the day and most days of the week
throughout the year. The users stay online, but their transactions may need restarting, and they may
experience performance degradation.
AE-2 (high availability)AE-2 covers business functions that allow minimally interrupted
computing services, either during essential periods or during most hours of the day and most days of
the week throughout the year. The users may be interrupted but can quickly log back in. They may
have to rerun some transactions from journal files, and they may experience performance
degradation.
AE-1 (high reliability)AE-1 covers business functions that can be interrupted as long as the
integrity of the data is insured. The user work stops and an uncontrolled shutdown occurs, but the
data integrity is ensured.
AE-0 (conventional)AE-0 covers business functions that can be interrupted and environments
where data integrity is not essential. The user work stops and uncontrolled shutdown occurs, and
data may be lost or corrupted.

Achieving availability
High levels of availability are not only expected but also required. Designers, engineers, and
manufacturers build resilience, availability, and fault tolerance into components, subsystems, and entire
solutions depending on whats required and what the customer is willing to pay for. Such solutions are
often built around these three goals:

1.Reduce or eliminate single points of failure by introducing and employing redundancy. This
means that a failure of a single component does not cause the entire system to fail.
2.Include a failover mechanism from one component to its redundant counterpart. Depending on the
solution design, the failover may be near instantaneous to taking a few seconds or minutes.
3.Provide adequate monitoring and failure detection, which then leads to quick recovery.
As you might suspect, protection against MTBF-/MTTR-related problems plays a key role in achieving
high levels of availability. To determine which highly available solution is right for you, you should
1.Determine the desired or the required level of availability.
2.Determine the cost of downtime.
3.Understand the recovery process, time, and procedures.
4.Focus on events that have negative aspects (events that can bring your system down as well as
events that may prevent or delay its recovery).
5.Develop a cost-consideration model, which weighs the cost of downtime, as defined earlier in
this chapter, versus the cost of the high-availability solution and alternatives.
6.Obtain the necessary approvals, funding, and support and then design and implement the solution.
For a distributed system, if a particular part of the system is unavailable due to a communication or other
failure, the system as a whole may still be deemed to be in the up state even though certain applications
may be down. When running multiple applications, the system may be fully functional and available even
if certain parts of the system are down.

Note
The type of availability at the high end of the availability scale is called continuous availability
or nonstop computing and is around 99.999% ((five nines)) of uptime or higher. This level of
availability yields 5.26 minutes of downtime per year (25.9 seconds per month, 6.05 seconds per
week, or 864.3 milliseconds per day), which is often not necessary. Many businesses are
satisfied with the requirement that the system does not go down during normal business hours.

Note
For additional details regarding high availability definition, requirements, and parameters, see:
https://en.wikipedia.org/wiki/High_availability.

RPO and RTO


Disaster recovery specialists examine the impact possibilities and the needs for availability in terms of
recovery point (also referred to as RPO) and recovery time (also referred to as RTO; Figure 1.7).

Figure 1-7 RPO vs. RTO

Recovery Point Objective


RPO is the acceptable amount of data loss, specified as the maximum % of data or the maximum length of
time, due to an event or an incident. This is the point or state of data to which the recovery process of a
given solution must be capable of returning. In terms of backups, the recovery point provides you with
information necessary to determine the frequency of performing backup operations.
RPO provides answers to these questions:
How much data is acceptable to be unprotected?
To what point in time do you need to recover (e.g., 24 hours, 1 hour, or several seconds)?

Recovery Time Objective


RTO defines the length of time necessary to restore the system operations after a disaster or a disruption
of service. RTO includes at minimum the time to correct the situation ((break fix)) and to restore any
data. It can, however, include factors such as detection, troubleshooting, testing, and communication to the
users.
Recovery = environment break fix + restore from an alternate source (backup disk or tape)

RTO provides answers to these questions:


How long is the customer or user willing to wait for the data or application?
How long can you tolerate downtime?

Note
Both RPO and RTO represent recovery objectives, or goals. These are not mandates and are
usually not treated as absolute requirements. Most disaster recovery specialists and businesses
select availability and recovery tactics that do not necessarily (or always) meet RPO and RTO,
but get close.
The availability requirements of a given solution or environment dictate the RTO and provide the basis
for determining these backup parameters:
The type of backup
The number of restores that are necessary

The location of the backup data


The speed and type of access to the backup data
Here are a few questions that you should ask your customers (and collect answers for) when defining the
appropriate protection and recovery strategy:
In terms of the operational impact of downtime:
o What is more important, a fast recovery or a recovery to an exact state prior to the failure? Or are
both important?
o What is the impact on operations relative to the recovery point? If you do not resume where you
left off, will the loss be inconvenient, damaging, or catastrophic?
o _________________________________________________________________ (add your
own)
o _________________________________________________________________ (add your
own)
o _________________________________________________________________ (add your
own)
In terms of the business impact of downtime:
o What is the business impact of the recovery time? If you do not resume processing within
seconds/minutes/hours/days, will it be inconvenient, damaging, or catastrophic to the business?
o What is the most effective and efficient method to recover your information?
o _________________________________________________________________ (add your
own)
o _________________________________________________________________ (add your
own)
o _________________________________________________________________ (add your
own)
Careful and complete assessment of the RPO and the RTO helps you define the protection and recovery
strategy, and which availability technologies, products, and solutions you recommend and use.

Availability technologies
As you can imagine, the list of availability approaches, technologies, products, and solutions is long.
There are hardware solutions, software solutions, and a combination of both. There is replication,
clustering, snapshots, clones, and RAID, to name just a few. Some protect against single points of failure,
others protect against magnitudes of incidents and disasters.
Table 1.2 provides a quick overview of some options that may be used to achieve the required levels of
availability.

Table 1-2 Options that may be used to achieve availability


To protect against:

Technology which can be used:

Power outages

UPS, generators

Component failures

Fault tolerance, RAID

Failure of the entire system, operating system, or interconnects

Fault tolerance, clusters

Administrative, planned downtime

Clusters

Regional or metropolitan downtimes

Remote copy

Redundant Array of Inexpensive (or Independent) Disks


The first priority when working with data is to protect it from disk failures. Considering the high price of
hard drives a few decades ago, companies struggled with two basic problemsdisk reliability and disk
size (or the available storage space). As a result, technology called Redundant Array of Inexpensive (or
Independent) Disks (RAID) was created.
The purpose of RAID is to combine smaller disk drives into larger ones or to introduce redundancy into a
system to protect it against disk drive failures or both. Several types of RAID configurations were
created, called RAID levels. These are most common:
RAID 0 (striping)is used to create larger virtual disks from smaller physical ones. Data blocks
are spread (striped) across these physical member disks with no redundancy. RAID 0 requires at
least two physical disk drives. An advantage of RAID 0 (Figure 1.8) is its performanceit
increases the number of spindles used to service I/O requests. However, due to no data protection,
you should never use this RAID in a business-critical system.

Figure 1-8 RAID 0

RAID 1 (mirroring)is used to protect the data by making writes to two identical disk drives at the
same time, thus keeping an exact copy in case of a disk failure. RAID 1 (Figure 1.9) requires at least
two physical disk drives and requires them in an even quantity. It provides good performance and
redundancy (it can withstand failure of multiple disk drives, as long as they are not mirrored to each

other). Its drawback is that the usable disk space is reduced by 50% due to information redundancy.

Figure 1-9 RAID 1

RAID 5 (striping with parity)was created as a cheaper alternative to mirroring, requiring fewer
disk drives in the array and being able to sustain a loss of a single disk drive without losing data.
RAID 5 requires at least three disk drives, provides good performance thanks to its data blocks
being striped, and good redundancy (using distributed parity). It also provides 67%93% of usable
disk storage. RAID 5 (Figure 1.10) is a great, cost-effective choice for predominantly readintensive environments, but suffers from slow write operations.

Figure 1-10 RAID 5

RAID 6 (striping with double parity)is similar to RAID 5, but it can sustain a loss of two disk
drives while keeping the data safe. It uses block-level striping for good performance and creates
two parity blocks for each data stripe. It requires at least four disk drives. RAID 6 (Figure 1.11) is
also more complicated to implement in the RAID controller or in a software.

Figure 1-11 RAID 6

RAID 10 (combining mirroring and striping)is a hybrid of RAID 1 and 0. It protects data by
maintaining a mirror on the secondary set of disk drives while using striping across each set to
improve data transfers. RAID 10 (Figure 1.12) requires at least four disk drives and an even number
of them. It provides excellent performance and redundancy, which is ideal for mission-critical
applications, at the cost of losing 50% of usable disk drive space.

Figure 1-12 RAID 10

Note
For
more
information
about
standard
https://en.wikipedia.org/wiki/Standard_RAID_levels.

RAID

levels,

go

to:

Important
RAID technology is necessary for system uptime and data protection against disk drive failure,
but it does not replace backups. It does not provide a point-in-time copy of the data, so the
information is not protected in case of a major hardware failure, user error, logical corruption,
viruses, or cyberattacks. It is also not designed to keep offline or offsite copies of your data,
which is the primary purpose of backups.

Snapshots
A snapshot is a point-in-time copy of data created from a set of markers pointing to stored data. Snapshots
became possible with advances in the Storage Area Networking (SAN) technologies, and they
complement the traditional RAID technology by adding backup functionality.
Most of the snapshot implementations use a technique called copy-on-write. The idea is to make an initial
snapshot and then further update it as data changes. Restoring to a specific point in time is possible as
long as all iterations of the data are kept. For that reason, snapshots can protect data against corruption
(unlike replication, which cannot).
Another common snapshot variant is the clone, or the split-mirror, where reference pointers are made to
the entire content of a mirrored set of drives, a file system, or a LUN every time the snapshot is created.
Clones take longer to create as compared to the copy-on-write snapshots because all data is physically
copied when the clone is created. There is also an impact on production performance when the clone is
created since the copy process has to access the primary data at the same time as the host.

Note
For
more
information
about
https://en.wikipedia.org/wiki/Snapshot_(computer_storage).

snapshots,

go

to:

Replication
Replication is a technique which involves making a point-in-time copy of data on a storage device, which
can be physically located in a remote location. Replication is sensitive to RTO and RPO and includes
synchronous and asynchronous replication, where the data transfer to the remote copy is achieved either
immediately or with a short time delay. Both methods create a secondary copy of the data that is identical
to the primary copy, with synchronous replication solutions achieving it in real time.
With replication, any data corruption or user file deletion is immediately (or very quickly) replicated to
the secondary copy, therefore, making it ineffective as a backup method.
Another point to remember with replication is that only one copy of the data is kept at the secondary
location. The replicated copy does not include historical versions of the data from preceding days, weeks,
or months, as is the case with backup.

Note
For

more

information

about

replication,

go

to:

https://en.wikipedia.org/wiki/Replication_(computing).

Calculating systems availability


To calculate the overall availability of a system, you must know the systems entire MTBF figure (based
on the MTBFs individual components) and the MTTR. Then, you can use this formula:
Availability =

Planning for availability


It is often challenging to determine which level of availability is best for a given business environment,
for these two reasons:
1. It is difficult to estimate the actual levels of availability that are required to meet given service levels
while also meeting budgetary and implementation scheduling requirements.
2. Even if you are successful with the initial estimates, the system usage patterns often change and shift
from one area to another, distorting the original assumptions and parameters.
When planning for availability, consider seeking answers to these questions:

Who are the customers or the users of the system and what are their expectations and requirements?
How much downtime are they willing to accept?
How much uptime are they willing to pay for?
What and who depends on the service your system provides?
What is the impact of downtime?
How often do you expect the system to be down?
How quickly can you recover from failures?
What are the alternatives and their cost/benefit analysis?
What is the budget and implementation timeframe?
What is the required skillset for its implementation, operation, and maintenance?
_____________________________________ (add your own)
_____________________________________ (add your own)
_____________________________________ (add your own)

Planning and designing for a specific availability level do pose challenges and tradeoffs that you must
make. There are a wide selection of technologies, products, and solutions on the market today, each with a
different methodology, architecture, requirements, benefits, and cost. No single availability solution fits
every situation. Therefore, a good solution always depends on a combination of business requirements,
application-specific parameters, implementation timetable, and available budget.

Learning check questions


Reinforce your knowledge and understanding of the topics just covered by completing this learning check:
1. Match the availability level with its correct description:

2. What is a typical expected downtime of a fault-resilient system?


a)None
b)1 hour per year
c)8.5 hours per year
d)3.5 days per year
3. Which RTO goal should be expected from a highly available system?
a)Less than one second
b)Seconds to minutes
c)Minutes to hours
d)Hours to days
4. What does the following formula calculate?

a)RTO
b)RPO
c)System availability
d)Expected downtime

5. Which RAID levels can withstand a failure of two disk drives? (Select all that apply.)
a)RAID 0
b)RAID 1
c)RAID 5
d)RAID 6
e)RAID 10
6. Which data protection technique creates reference pointers to the entire content of duplicate drives,
file system, or LUNs every time a snapshot is made?
a)Archiving
b)Copy-on-write
c)Replication
d)Split-mirror
e)Deduplication

Learning check answers


This section contains answers to the learning check questions.
1. Match the availability level with its correct description:

2. What is a typical expected downtime of a fault-resilient system?


a)None
b)1 hour per year
c)8.5 hours per year

d)3.5 days per year


3. Which RTO goal should be expected from a highly available system?
a)Less than one second
b)Seconds to minutes
c)Minutes to hours
d)Hours to days
4. What does the following formula calculate?

a)RTO
b)RPO
c)System availability
d)Expected downtime
5. Which RAID levels can withstand a failure of two disk drives? (Select all that apply.)
a)RAID 0
b)RAID 1
c)RAID 5
d)RAID 6
e)RAID 10
6. Which data protection technique creates reference pointers to the entire content of duplicate drives,
file system, or LUNs every time a snapshot is made?
a)Archiving
b)Copy-on-write
c)Replication
d)Split-mirror
e)Deduplication

Summary
There are many human errors, software glitches, hardware failures, and human-induced or natural
disasters, all affecting IT operations. Downtime due to these disruptions can result in lost productivity,
revenue, market share, customers, employee satisfaction, and in some extreme situations, even in loss of
life.
MTBFs and MTTR are two metrics that deal with downtime prediction and recovery.
Availability is the opposite of downtime. It measures the time during which a system runs and delivers
data, applications, or services. System availability varies from conventional systems to fully fault-tolerant
systems. So does their cost.
Backup and RAID technologies are not the same, and each has a different purpose. While both protect

your data, backups provide secondary copies of your data in case it is lost or corrupted, whereas RAID
(except for RAID 0) provides some level of redundancy in case of a disk drive failure.
Various RAID technologies exist, ranging from RAID 0 to RAID 10, each with different levels of
protection and performance characteristics.
RPO and RTO define the boundaries under which a system and its data are brought back into service after
downtime.
Redundant and fault-tolerant components address high-availability demands. Backup protects your data
against multiple hardware-component failures, human errors, viruses, cyber-attacks, and data corruption.

2 Backup Strategies

CHAPTER OBJECTIVES
In this chapter, you will learn to:
Define a backup, explain its purpose, and describe the basic backup terminology
List and describe backup types and explain which backup type is appropriate for a given situation
List and describe the most common backup tape rotation schemes and related topics such as the
overwrite protection and append period
Explain the restore operations
Define archiving and contrast it with backing up
Describe data tiering and its benefits
Explain the purpose of a backup and recovery plan and define the recommended steps in a sample
backup and recovery plan
List and describe common tape media drives and technologies

Introduction
In the previous chapter, you learned about downtime and availability. You learned about what causes
downtime, what impact it has on IT operations and on the business, and how it translates to cost. Then you
also learned about availabilityhow to achieve it, how to calculate it, and how to plan for it.
The previous chapter also made a clear distinction between availability technologies, such as RAID and
replication, and explained why backups are critical in protecting business applications and data.
This chapter focuses on backup strategies and related topics such as restore, archiving, and data tiering.
As you may already know or suspect, backup is a wide topic, which must be understood to properly plan
your backup strategies and fit into your overall protection implementation. Why? Because your backup is
as good as the data you can restore from it.

Learner Assessment
Before continuing with this chapter, take a few minutes to assess what you might already know by defining
these topics:

Note
The purpose of this exercise is to get you thinking about these topics. Do not worry if you do not
know the answers; just put in your best effort. Each topic will be explained later in this chapter.
In your own terms, define backup and its purpose:
___________________________________________________________
___________________________________________________________

___________________________________________________________
___________________________________________________________
___________________________________________________________
___________________________________________________________
___________________________________________________________
In your own terms, define these concepts:
Backup consistency
___________________________________________________________
___________________________________________________________
___________________________________________________________
___________________________________________________________
Backup application
___________________________________________________________
___________________________________________________________
___________________________________________________________
___________________________________________________________
Backup device
___________________________________________________________
___________________________________________________________
___________________________________________________________
___________________________________________________________
Deduplication
___________________________________________________________
___________________________________________________________
___________________________________________________________
___________________________________________________________

What is a backup?
Now, consider these questions regarding backup:

Is backup a protection against hardware failures?


Does backup address disk-related problems?
Is backup the best protection against loss of data due to hardware failures?
Does backup always have the latest version of the data?
Is backup still vulnerable to some data loss?

Everyone would agree that a backup is an insurance policy against data loss. A company depends on its
information to stay in business, whether the information takes form of email messages, a customer
database, research and development projects, or even unstructured data. An individual depends on his/her

information for work-related tasks. Even in our daily personal lives, we rely on information which we
cannot (or do not want to) lose.

Backup definition

Figure 2-1 Backup

A backup (Figure 2.1) is the process of creating a copy of data, applications, or operating systems on a
secondary storage medium. This copy is created, stored, and kept for future use in case the original is
destroyed or corrupted.
According to the Dictionary of Computer Terms:
(Backup is the activity of copying files or databases so that they will be preserved in case of equipment
failure or other catastrophes. Backup is usually a routine part of the operation of large businesses with
servers as well as the administrators of smaller business companies. For personal computer users, backup
is also necessary but often neglected.)
In most cases, the source is data on a disk, such as files, directories, databases, and applications. The
target is another disk or a tape, hosted by some type of a tape drive or a tape library.
Since the backup is expected to be used for disaster recovery, it must be consistent. Consistency means
that the data copy is the same as the original, and when restored, it is identical. Database consistency is
achieved by consistency checks such as byte-by-byte comparison, cyclic redundancy checks, or reduced
verification.
The software that copies the data to its destination is called the backup application. The destination is
called the backup device, such as a tape drive, with a medium onto which the copy of the data is written.
The backup device can be physical or virtual. There are many benefits of virtualizing the target backup
device, one of the most significant being the potential to take advantage of the deduplication functionality,
and another to overcome the limitations of sequential physical devices, such as speed and size.
A customers or your own company recovery requirements help you decide on the best way to back up the
data, and there is rarely a case where all data should be treated the same way. This means that retention
requirements (how long to keep the backup) vary for different types of data.

Backup purpose
As already mentioned, backup is an insurance policy against data loss. It plays a critical role in
Disaster recovery (enables data to be restored after a disaster)

Data archival (consists of files or records that have been selected for permanent or long-term
preservation)
Operational backup (enables a restore of a small or a selected number of objects, such as files or
emails, after they have been accidentally deleted or corrupted)
Recovery from data corruption (such as databases)
Compliance with corporate guidelines and/or local, state, and federal requirements for data retention

Backup sets
What should a backup consist of?
There is no rule on what must be backed up and what can be left unprotected. Some of the good
candidates to be included in the backup sets are the following:
Operating environments
o Laptop operating systems
o Desktop operating systems
o Server operating systems (physical and virtual)
Applications
o Enterprise Resource Planning (ERP) applications such as SAP, Oracle Applications, and
PeopleSoft
o Customer Relationship Management (CRM) applications such as SAP Digital, Siebel, Salesforce,
and Microsoft Dynamics CRM
o Databases, such as Oracle, UDP, and Microsoft SQL Server
o Messaging applications, such as Microsoft Exchange
User and application data (for all above operating environments and applications)
Logs and journals
o Application transaction logs
o Database journals
o File system journals
Always consider including in your backup sets everything that you cannot afford to lose.

Backup types
This section defines the archive bit, which is an important property of operating system files and plays a
key role in performing backups, and lists and describes various types of backups.

Archive bit
When a file is created or changed, the operating system maintains a flag, which is called the archive bit.
The backup software uses this bit to determine whether the file has been backed up before. As soon as the
file is backed up using either the full or incremental backup, this bit is turned off, indicating to the system
that the file was saved to a secondary medium. If the file is changed again, the bit is turned back on, and

the file is flagged to be backed up again by the next full or incremental backup. Differential backups
include only files that were created or modified since the last full backup. When a differential backup is
performed, no changes are made to the archive bit.

Archive bit

ON

OFF

File must be backed up

File has been backed up

A copy is an option in many backup applications. Its function is similar to a full backup; however, the
archive bit is not modified.

Learner assessment
You may already be familiar with many types of backups. For fun and to test your preexisting knowledge,
see if you can answer these questions:

Note
The purpose of this exercise is for you to assess what you already know. Do not worry if you do
not know the answers; just put in your best effort. Each topic will be explained later in this
chapter.
Which backup type is described by these statements?
It is also called a normal backup.
It saves the entire content of the source disk on a file-by-file basis.
After backing up individual files, it turns the archive bit off.
a)Online backup
b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup
Which backup type is described by these statements?
It enables restoring individual files to their original locations.
It provides random access to individual files for a quick restore.
It requires a full path to be saved in the backup set.
It can create significant operating system overhead with a large number of small files.
a)Online backup
b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup
Which backup type is described by these statements?

It supports 24x7 operations because it does not require a system shutdown.


It can create a degradation in server and network performance.
It may create data integrity and inconsistency issues.
a)Online backup
b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup
Which backup type is described by these statements?
It saves the entire disk at the block level.
It is often called physical backup because it dumps the complete file system to a single file.
It provides the fastest method to save or to recover a complete file system
It requires a restore to identical disk.
a)User-defined backup
b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup
Which backup type is described by these statements?
It is performed upon a special request, which defines a set of files to back up.
It is performed in addition to and outside of the standard tape rotation scheme.
It is usually performed before a system or software upgrade or when necessary to save certain
files.
a)User-defined backup
b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup

Learner Assessment Answers


This section contains answers to this chapters learner assessment questions.
Which backup type is described by these statements?
It is also called a normal backup.
It saves the entire content of the source disk on a file-by-file basis.
After backing up individual files, it turns the archive bit off.
a)Online backup
b)Offline backup
c)File-based backup

d)Image-based backup
e)Full backup
Which backup type is described by these statements?
It enables restoring individual files to their original locations.
It provides random access to individual files for a quick restore.
It requires a full path to be saved in the backup set.
It can create significant operating system overhead with a large number of small files.
a)Online backup
b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup
Which backup type is described by these statements?
It supports 24x7 operations because it does not require a system shutdown.
It can create a degradation in server and network performance.
It may create data integrity and inconsistency issues.
a)Online backup
b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup
Which backup type is described by these statements?
It saves the entire disk at the block level.
It is often called physical backup because it dumps the complete file system to a single file.
It provides the fastest method to save or recover a complete file system
It requires a restore to identical disk.
a)User-defined backup
b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup
Which backup type is described by these statements?
It is performed upon a special request, which defines a set of files to back up.
It is performed in addition to and outside of the standard tape rotation scheme.
It is usually performed before a system or a software upgrade or when necessary to save certain
files.
a)User-defined backup

b)Offline backup
c)File-based backup
d)Image-based backup
e)Full backup

Offline backup
For an offline backup, the administrator takes the applications and their data offline for the time required
to do the backup. During the backup window, the host (such as a server or a storage array) is not
available to its users, which is why this type of backup is usually done at times when the user demand is
at its lowest (such as at night or on the weekends).
The exact calculation of the overall system performance and the volume of backup data are both required
to determine whether the actual time needed for the operation fits into the allotted backup window to
avoid extended offline periods for the host.
From the backup perspective, offline backup is the easiest and most secure form of performing backups,
because the backup software usually does not have to deal with open files or performance issues. An
offline backup can be complete or partial.

Online backup
As more companies move toward 24-hour and 7-days-per-week operations, no clear backup window for
offline backups exists. Online backup is the alternative and is performed during normal operating hours
and with the host fully available to its users. Usually both the users and the backup environment see a
degradation in the overall network and host performance during the backup window, thus influencing
productivity. Because user and system files can be open during online backups, there is a danger to data
integrity, which has to be solved appropriately by the backup software.
Inconsistent files could mean a serious threat to the data integrity when restored from the backup medium.
Incompletely backed up files (open files with ongoing write operations while the backup software is
copying them to tape) may contain inconsistent data which could be spread across various databases in
the enterprise, depending on how the databases are linked to each other. An example of inconsistency is
database indices that no longer reflect the actual content of the database.
Your backup software should at least offer the option of selecting what to do with open fileseither copy
them anyway, mark the files in a log file as suspect, or do not copy these files but inform the administrator
via a log file or an alert. Some backup programs automatically retry open files after a certain time, when
there may be a chance that the file has been closed in the meantime. An online backup can also be
complete or partial.

File-by-file backup (file-based backup)


With file-by-file backups, the information needed to retrieve a single file from the backup medium is
retained and it is possible to restore files to their correct locations on the disk. The full path information
for every file must be saved on the backup medium, which, in an environment with a large number of
small files, could cause a significant overhead for the operating system. The operating system must access
the file allocation table many times and can reduce the backup throughput by as much as 90%. The
performance can also be influenced by other factors such as disk fragmentation. However, file-by-file

backup provides random access to individual files for a quick restore.


Furthermore, this type of backup is used by all tape rotation schemes to back up selected files.

Image backup (image-based backup)


An image backup:
Saves the entire disk at the block level.
Provides the highest backup performance due to large sequential I/Os (the tape drive is typically
kept streaming).
Creates a snapshot of the disk at the time of the backup.
Requires a restore to an identical disk.
Is used for archiving because it does not restore individual files.
Image backups may also be referred to as physical backups and are used to describe a dump of the
complete file system to a single backup image file. Consequently, only the complete image can be
restored.
When the image backup is performed locally, the tape drive typically keeps streaming. If the image
backup is done over the network to a distant tape drive, other activities on the network may influence the
tape drives performance.
An image backup is usually the fastest method to save or to recover a complete file system. In
environments with an exceptionally high number of small files, this could be your best choice of backup
methods.

User-defined backup
A user-defined backup is

Performed upon a special request from a user who defines which set of files to back up
Performed in addition to and outside of the standard tape rotation scheme
Defined by the backup set which includes user files
Performed before a system or a software upgrade

A user-defined backup usually means a special backup request where the user defines which files need to
be copied (e.g., for a critical project). This can be the complete user account with all applications,
configurations, and customizations, which can be especially useful when the user plans to do a major
system change like operating system or software upgrades. Another request could consist of just the
information base of the user, which includes certain applications and their data, or all information created
only within a certain timeframe.

Full backup
A full backup is also called a complete or normal backup. It saves the entire content of the source disk to
the backup medium on a file-by-file basis. In networked environments, this source disk could be the disk

of a client machine. After backing up individual files, their archive bit is turned off.
A full backup is time-consuming but critical to a full disaster recovery. Therefore, you should always start
with a full backup.

Differential backup

Figure 2-2 Differential backup

A differential backup (Figure 2.2) copies all files which have been modified or created since the last full
backup, but does not modify the archive bit.
A differential backup is useful when you want to have the latest version of files on a tape in the latest
backup set. When the same tape is used for the differential backups between full backups, usually the
newer file versions are allowed to overwrite the older versions of the same file. Special caution must be
taken when database log files are backed upcircular logging must be disabled and the log files must not
be deleted after the backup.
Since the archive bit of the files does not change during the differential backup, the files appear to the
system as not being backed up and are saved again with the next differential backup.

Differential backupdata to be saved calculation and associated example

Data to be saved:
B = base amount of data
x = daily changes in %
o = data overlap in %
t = number of days

= data to be saved on tape after t days of DIF backup

Example 1
Here is one example:
B = 200 MB data on disk (first full backup)
x = 10% daily changes
o = 90% daily overlap
t = 5 days

Example 2
As a different example (from the above graphic), imagine that you have
1 GB of data on the entire disk, which is backed up with a full backup on Monday onto its own tape.

50 MB of data that changes daily (there is a separate tape available for each day).
Then with a differential backup:
Tuesday tape contains 50 MB.
Wednesday tape contains 100 MB
Thursday tape contains 150 MB
If a complete data loss occurred on Friday, the restore operation would only need:
The full backup tape, which was created on Monday
The last good differential backup tape (the one created on Thursday)
The advantage of a differential backup is that only two tapes are necessary to completely restore the
system (the last full backup tape and the last differential backup tape). The disadvantages are the volume
of data on these tapes, which grows every day, the time for the backup, which also increases every day,
and only the latest versions of files can be restored.

Incremental backup

Figure 2-3 Incremental backup

An incremental backup (Figure 2.3) copies all files which have been modified or created since the last
full or incremental backup and turns the archive bit off.
With incremental backups, the complete history of the files is maintained. All files that have been created
or modified since the last full or incremental backup are saved. Each revision or version of the files is
available on the tape because the backup software does not overwrite the earlier versions of the same
files but instead appends the data to the previous backup sets (or uses different tapes for each incremental
backup).
When backing up database log files, circular logging must be disabled. Like the normal backups, the
incremental backup purges the log files after the backup process.
The formulas to calculate the size of data to be saved on a certain day and the total size of data after a
number of days to be restored are

B = the initial data size (the first full backup)


x = daily changes in %
g = daily growth in %
t = number of days
At = data to be saved on day t with incremental backups

Tt = total size of data which must be restored after t days of incremental backups
Example
Imagine that you have
1 GB of data on the entire disk, which is backed up with a full backup on Monday onto its own tape.
50 MB of data changes daily (there is a separate tape available for each day).
Then with an incremental backup:
Tuesday tape contains 50 MB
Wednesday tape contains 50 MB
Thursday tape contains 50 MB
If a complete data loss occurred on Friday, the restore operation would need
The full backup tape, which was created on Monday
Every incremental backup tape (Tuesday, Wednesday, and Thursday)
The advantages of the incremental backup are that the size of the backed up data on each tape remains
almost the same, and the backup window is small each day. The disadvantage is that each day must be
restored separately, thus consuming time.

When to choose incremental vs. differential backup


Choose incremental backup when
You have a high percentage of daily changes in your data set.
You only have a limited backup window per day.
You can accept a relatively slow restore (which translates to a long downtime).
Choose differential backup when
You have a low percentage of daily changes in your data set or the same files changing every day.

The backup window is not an issue (you have enough time available for the backup).
You need a relatively fast restore (which translates to a short downtime).
Selecting the appropriate type of backup (incremental or differential) is a trade-off between the time
needed for the backup and the time needed for the restore, which depends on the size of data and the daily
change rate. The type of backup must be considered when selecting the right tape rotation scheme.
The formula for calculating the data growth rate is
Tt =

B = initial data size on the disk (on the first day of calculation)
g = growth rate in % of B per time unit
t = time unit (days, weeks, months, ...)
Tt = total data size after t

Example
Imagine that you have
B = 200 MB of data on the disk
g = 2% daily growth
How much data do you have after one year (t = 365)?
T365 = 200 MB

Backup to disk
Backups are traditionally connected with tape drives. Tape drives, despite their advances in speed and in
capacity, do have disadvantages. One of the biggest disadvantages is their serial nature. For any bit of
information located at the end of the tape, the preceding tape medium must be accessed in order for that
information to be accessed. This does not pose a problem for large chunks of information, but when you
must restore small amounts of data, such as a single mailbox or a file, the need for a different approach
becomes obvious.
Backup to disk (B2D) is a method of backing up large volumes of data to a disk storage unit. This method
presents benefits such as
Backup to disk is very fast. In fact, the speed of B2D would likely be limited by the hosts
performance capabilities rather than the infrastructure or the backup destination speed, since the
B2D systems can use a dedicated iSCSI or Fibre Channel infrastructure and SSD drives to store the
information.
HPE offers the B2D capabilities either through a direct backup to disk devices implemented by the
Data Protector or by using dedicated B2D storage devices such as HPE StoreOnce, which provide
much more functionality than a simple backup to a disk target.

B2D allows for data deduplication at the source or at the target.


B2D allows for staged backup scenarios, where data is first backed up to a disk, and then, after the
peek hours, moved to the tape. This type of backup is called backup-to-disk-to-tape. Such an
environment allows for better speed, since the backup to disk can be almost instantaneous, thus
improving the RTO and RPO times.
B2D allows for data replication to a remote site. This is particularly efficient when combined with
deduplication so that only deduplicated data flows through the network.

Synthetic backup
During the over-the-network backup, data could be moved from many clients to the central backup server,
and then moved to a permanent storage such as a tape library. Moving data around the data center is
always expensive and any operation such as creating a full backup of many clients can overload the
network infrastructure.
Considering that all backup sets (full, incremental, and differential) end up on a single backup server, it is
possible to take the last full backup and append the incremental or differential backups to the full backup
set. Once the appending procedure is done, the result is stored to the same backup target as the source
data sets and is promoted to the full backup. This methodology is called synthetic backup.
The main benefit of synthetic backups over classic full backups is that it does not involve the clients or
moving the information over the network. If planned properly and executed at the right backup server at
the right time, synthetic backup can be created almost instantly.
Synthetic backups do not provide any advantages over the regular full backups when performing the
restore procedures.

Data Protector virtual full backup


HPE Data Protector provides the virtual full backup option as an efficient type of synthetic backup where
data is consolidated using pointers instead of being copied. It is performed if all backups (the full backup,
incremental backups, and the resulting virtual full backup) are written to a single file library that uses the
distributed file medium format. The distributed file medium format is a media format available with the
file library, which is a prerequisite for synthetic or virtual full backup functionality.

Working set backup


A working set backup is similar to a differential backup; however, the user specifies a range of time and
only data that has been created or modified since the last full or incremental backup is backed up, plus the
data that has been accessed ((touched)) during the specified time frame.
Working set backups can speed up the recovery of a crashed server because you only need to restore the
working set backup to get the server up and running again and then restore the latest full backup at a later
time (if necessary).
The advantages of using the working set backup are the following:
Restoring a system backed up with a working set strategy requires only the medium containing the
latest working set backup and the most recent full backup.

You can perform a working set backup, restore the data to a new system, and have the system up and
running faster as compared to restoring a full backup and all incremental or differential backups.
Working set backup takes less time to run than full backups.
The disadvantage is that along with all files accessed during the specified time, all files created or
modified since the last normal or incremental backup are included on each medium, thus creating
redundant working set backups.

Block-level incremental backup


A typical database, such as Oracle or Microsoft SQL Server, consists of a small number of relatively
large container files. Inside those files, there is usually only a small fraction of real user data changes
between backups. Thus, a standard incremental backup copies the entire container file even though only a
small percentage of the data changed (effectively becoming a full backup with a resulting high usage of
tape space and generated I/Os). With Block Level Incremental Backup (BLIB), only those database blocks
that have been changed within the container file are recorded in a changed block map.
Backup applications with the BLIB functionality, such as Veritas NetBackup, copy only the changed
database blocks based on the information in the changed block map and typically take very little time and
use only small percentage of backup medium storage and I/O bandwidth.
A full database restore requires the last full backup of the database and all block-level incremental
backup sets since then.
Usually, the BLIB functionality is part of the backup applications database agent.
BLIB increases data availability because
It can reduce or eliminate the need for backup windows.
It allows for more frequent backup schedules.
The backup sets contain up-to-date data.

Snapshots (frozen images, off-host backups, or nondisruptive backups)


The need to run backups without stopping the applications led to the development of the frozen image
technology or snapshots. With snapshots, an image of the online data is frozen. A snapshot is a record of
blocks of data at a particular instant. The image is backed up while the application continues to update the
main data.
There are two popular snapshot techniques:
Breakaway mirrors (clones)
At least two identical sets of data are maintained on separate disks. Before a backup begins, all
database activity is quiesced (paused) to ensure that buffers and cached data are flushed to disk.
Then one of the disks (the clone) is split off and remounted in the read-only mode and then backed
up. During the backup, the application can update the other disk.
Breakaway mirrors are reliable, but they require at least two sets of data. They require additional
storage and generate I/O traffic needed to synchronize the disks after backup. To reduce these
disadvantages, copy-on-write was developed.

Copy-on-write (point-in-time)
To create the snapshot, all access to the data is blocked (the application is quiesced), and the cache
is flushed to disk to make the data consistent. All changes to the data are recorded in a changed
block map (these actions take only a few seconds at a certain point in time (X), after which the
application can resume its operations).
The snapshot itself is then mounted as a read-only file system. When the application writes the data,
the original data being overwritten is first copied to the snapshot and the changed block map is
updated. When the backup (e.g., at the time of X + 10 minutes) reads the data, it is delivered from
the original location, if it has not been modified, or from the snapshot (at the time X), if the original
has been modified. The snapshot insures point-in-time data integrity.

Backup tape rotation schemes


A proper backup strategy should attempt to
Maintain a file history by saving several versions over a period of time
Minimize the backup window by only saving what has changed
The solution for these tasks is a tape rotation scheme.
The three most popular tape rotation schemes, which are configurable with many backup applications, are
discussed in this section. In addition to the standard features, the tape rotation schemes provide older
copies of data on tapes, which can be archived offsite in a secure location to provide disaster tolerance.

What are backup tape rotation schemes?


A good plan is necessary to ensure that your backups are performed at the appropriate intervals and not
only when neededbecause data is valuable and difficult to replace. Most companies back up their data
on a daily basis when the network is least busy. These backups may need to be performed more or less
frequently, based on the criticality of the data. A regular and scheduled backup should address these
issues:
Speed of recovery after a disaster occurs
Redundancy and history of the data
Efficient use of tape media (although tapes are relatively inexpensive)
Automation helps to increase the overall speed of both backup and restore. The appropriate backup
schedule and the data version history can be automated with a tape rotation scheme. The longer a
company needs to keep its data, the more portable media (tape cartridges) are needed. The three biggest
advantages of tape rotation schemes are the following:
Automation
Archiving
File history
Working with tape drives requires planning. Considering a relatively limited capacity and the price of the

tape media compared to the volume of data being backed up, you want to reuse the tape cartridges by
rotating them. While using a single external tape drive, it is the administrators task to replace the tapes
and enforce the rotation order. Modern high-capacity tape libraries implement the tape rotation
themselves and the administrator selects the rotation scheme as an option within the backup server
settings.
The most popular backup tape rotation schemes are the following:
First-In, First-Out (FIFO)
Grandfatherfatherson
Tower of Hanoi

First-In, First-Out (FIFO)

Figure 2-4 FIFO

The FIFO (Figure 2.4) backup scheme saves new or modified files on the oldest medium in the backup set
(e.g., the medium which contains the oldest and thus least useful previously backed up data). Performing a
daily backup onto a set of 14 media, the backup depth would be 14 days. Each day, the oldest medium
would be inserted when performing the backup. This is the simplest rotation scheme and is usually the
first to come to mind.
This scheme has the advantage of retaining the longest possible tail of daily backups. It can be used when
the archived data is unimportant (or is retained separately from the short-term backup data) and any data
before the rotation period is irrelevant.
This scheme suffers from the possibility of data loss: suppose an error is introduced into the data but the
problem is not identified until several generations of backups and revisions have taken place. When the
error is detected, all backup files contain this error.

Grandfatherfatherson

Figure 2-5 The grandfatherfatherson

The grandfatherfatherson backup (Figure 2.5) refers to a common rotation scheme for the backup
media. In this scheme, there are three or more backup sets, such as daily, weekly, and monthly. The daily
backups are rotated on a daily basis using the FIFO scheme described earlier. The weekly backups are
similarly rotated, but on a weekly basis, and the monthly backups are rotated on a monthly basis. In
addition, quarterly, half-yearly, and/or annual backups could also be separately retained. Often, some of
these backups are removed from the site for safekeeping and disaster recovery purposes.

Tower of Hanoi

Figure 2-6 The Tower of Hanoi rotation method

The Tower of Hanoi rotation method (Figure 2.6) is the most complex rotation scheme to explain and use.
It is based on the mathematics of the Tower of Hanoi puzzle, using a recursive method to optimize the
backup cycle.

Note
There are several videos available on the Internet that visually explain the Tower of Hanoi
algorithm. One such video is the 4-minute 11-second video titled The Towers of Hanoi
Algorithm at https://www.youtube.com/watch?v=o8iTGJcEO-w.

Figure 2.7 represents the five tape rotation schedule, using tapes labeled as A, B, C, D, and E.

Figure 2-7 Five tape rotation schedule

In this schedule, one tape set A is used every other backup session (daily sessions 116 are shown in this
example across the top). Start day 1 with tape set A and repeat every other backup session (every other
day). The next tape set B starts on the first non-A backup day (day 2) and repeats every fourth backup
session (days 6, 10, and 14). Media set C starts on the first non-A or non-B backup day (day 4) and
repeats every eighth session (day 12). Media set D starts on the first non-A, non-B, or non-C backup day
(day 8) and repeats every sixteenth session (day 24, not shown). Media set E alternates with media set D.
The Tower of Hanoi rotation method implies that you keep only one backup per level (row in the above
table) and delete all outdated backups. The method allows for efficient data storage by having more
backups accumulating toward the present time. If you have four backups, you can recover data as of today,
yesterday, half a week ago, or a week ago. With five levels, you can recover data backed up two weeks
ago. Every additional backup level doubles the maximum rollback period for your data.
This schedule can be used in either a daily or a weekly rotation scheme. The decision regarding the
frequency of rotation should be based on the volume of data traffic. To maintain a reasonable history of
file versions, a minimum of five tape sets should be used in the weekly rotation schedule, or eight for a
daily rotation scheme. As with the grandfatherfatherson rotation scheme, tapes should be periodically
removed from the rotation for archival purposes.

Overwrite protection and append periods

Figure 2-8 Overwrite protection period

The overwrite protection period (OPP; Figure 2.8) defines the length of time to retain the data on the
medium before it can be overwritten. OPPs provide a foundation for all tape rotation schemes and are
defined per media or per media set. OPP is measured from the last time the data was written to the
medium.
The append period (APP) defines how long you are prepared to allow appending more jobs to a medium
once you began writing to it.

Example
In a weekly full backup, you use four media or media sets over a period of four weeks (one set for each
week). Thus, OPP is set to four weeks for each set. This protects the data from being accidentally
overwritten before the period is over. The same principle applies to daily incremental or differential
backup media (the protection period is set to one week, for instance).
A daily (incremental or differential) backup should be performed each day at the same time. The backup
starts at 1:00 and takes around 45 minutes to complete. You want to use the same tape every day and set
the backup application to overwrite.

Media rotation policies


Traditional backup strategies used with older backup applications required a thoroughly planned and well
defined media rotation policy controlled by the administrator rather than the backup application. Modern
backup tools, such as the HPE Data Protector, allow you to implement a rotation policy by specifying
usage options, such that media selection for subsequent backups.
A media rotation policy defines how media are used during the backup. The definition includes answers
to these questions:
How many backup generations are needed?
Where are the media stored?

How often are the media used?


When can the media be overwritten and reused for new backups?
When do the media become old enough to be replaced?
Media rotation is implemented with these characteristics:
Because the media are grouped into media pools, you no longer need to manage a single medium.
The software automatically tracks and manages every medium in the media pools.
You do not need to decide to which medium the backed up data is written to. You back up to a media
pool.
The software automatically selects the medium from the media pool according to the medium
allocation policy and usage options you specified. You can also disable the automatic selection and
perform a manual medium selection.
The medium location is tracked and displayed in the user interface once the medium is configured
with the backup software.
The software tracks the number of overwrites on the medium and the medium age, thus its condition.

Backup in a cloud
Cloud computing became widely accepted form of handling information and services. In early days of
cloud computing, companies had to deal with issues such as network speed, security, cost, and volume of
data hosted in the cloud. Little by little, these early issues have been addressed by infrastructure
improvements.

Figure 2-9 Cloud types

Based on the location of data, three models of cloud (Figure 2.9) computing exist
Public cloudruns beyond the company firewall and is usually hosted at a remote location operated
by a company specializing in cloud hosting (called the cloud hosting provider). The cloud hosting
provider makes the cloud resources, such as applications, compute, or storage, available to the

general public subscribers over the Internet. These cloud resources may be free or offered on a payper-use basis. Security, performance, and availability are the biggest concerns with public clouds.
Private cloudcontains data and services that are hosted behind a companys firewall. A private
cloud is dedicated to a single organization and although it often offers the power, efficiency, and
functionality of a public cloud, it features the security, control, and performance of a dedicated
computing environment.
Hybrid cloudincludes a mix of private and public hosting to get the best features from both the
private and public clouds.
The type of the cloud, or the cloud design, affects planning of data backups. With the private cloud, the
data is within the data center. With the public cloud, the data is at a third-party site. Many cloud service
providers offer some type of a remote backup, but the questions of the backup availability, location,
security, and compliance remain. In hybrid clouds, backups can experience issues and challenges related
to both the public and private clouds.

Learning check questions


Reinforce your knowledge and understanding of the topics just covered by completing this learning check:
1. Your customer has these backup requirements:
Low percentage of daily changes in the data set; some files change every day
Enough time available for backups
Fast restore time and short downtime requirement
Which type of backup should you recommend?
a)Full backup
b)Incremental backup
c)Image-based backup
d)Differential backup
2. What are the requirements to fully restore a database which was backed up with BLIB? (Select
two.)
a)Last full backup
b)Last block-level incremental backup
c)All block-level incremental backups since the full backup
d)Last block-level differential backup
e)All block-level differential backups since the full backup
3. Which backup types reset the archive bit associated with the operating system files? (Select two.)
a)Differential
b)Full
c)Incremental
d)Copy
e)Image

f)Block-level incremental
4. What does the block-level incremental backup use to determine what to back up?
a)The archive bit
b)The changed block map
c)The file allocation table
d)The database log file

Learning check answers


This section contains answers to this chapters learning check questions.
1. Your customer has these backup requirements:
Low percentage of daily changes in the data set; some files change every day
Enough time available for backups
Fast restore time and short downtime requirement
Which type of backup should you recommend?
a)Full backup
b)Incremental backup
c)Image-based backup
d)Differential backup
2. What are the requirements to fully restore a database which was backed up with BLIB? (Select
two.)
a)Last full backup
b)Last block-level incremental backup
c)All block-level incremental backups since the full backup
d)Last block-level differential backup
e)All block-level differential backups since the full backup
3. Which backup types reset the archive bit associated with the operating system files? (Select two.)
a)Differential
b)Full
c)Incremental
d)Copy
e)Image
f)Block-level incremental
4. What does the block-level incremental backup use to determine what to back up?
a)The archive bit
b)The changed block map

c)The file allocation table


d)The database log file

What is a restore?
This short section explains what a restore is and what possible challenges could be with restore
operations.

Restore definition

Figure 2-10 Restore process

Earlier in this chapter, the Dictionary of Computer Terms was used to define a backup. How is a restore
defined according to the Dictionary of Computer Terms?
(The retrieval of files you backed up is called restoring them.)
A restore is a process (Figure 2.10) that recreates the original data from a backup copy. This process
consists of the preparation and the actual restoration of the data, as well as any applicable postrestore
actions that make this data ready for use.
The source of the restore process is the backup copy. A restore application is the software that writes the
data to its destination. The destination is usually a disk to which the original data is written.
It was already mentioned that without a backup, you cannot restore your data. Both the backup and restore
operations must be part of your backup and restore strategy (your recovery plan). Many businesses do not
place enough emphasis on running the backup operations, other than that they must complete successfully
and not impact the normal business operations. Although you may be able to get away with achieving
these two primary goals during a backup, you must pay enough attention to your restore operations,
especially how quickly you can restore your data and return to production.

Possible problems with restore operations


When files are deleted after the last backup, they are no longer on the disk, but they are still part of the
last backup set on the backup medium. Eventually, the freed disk space could be used for new files,
especially if the disk usage was close to its capacity.
Unfortunately, with standard backups, there is no mechanism to track deleted files and keep their logs in a

journal. When you want to restore your backup sets (e.g., the last full backup and five incremental
backups created thereafter), you may end up restoring more data than available storage.
In this case, you can get an error message stating that your disk is full.
Some backup applications, such as Tivoli, Legato NetWorker, and Veritas NetBackup, have an option to
enable logging of deleted files. With this option, which must be enabled during the backup, it is possible
to skip files that have been intentionally deleted from the disk while restoring the backup sets.

What is archiving?
You may be wondering whether backing up and archiving are the same, or, if they are different, what their
differences are. This short section explains just that.

Differences between backing up and archiving


Depending on the goals of your data protection, you can decide to use archiving or backing up. Although
sometimes used together in the same context, backup and archiving have a different purpose. A backup is
used to keep copies of the data for data protection purposes, while archiving is done as a means of data
management for keeping the data organized for a long term.
In other words, a backup is used for short-term data protection and it might contain multiple instances of
the data, while archiving includes arranging and indexing the information in order to preserve the data for
a long time. Archives usually contain a single instance of the data.
You can delete the original data once it is archived because accessing this information immediately is
usually not required any longer. However, in reality, backed up data is usually not deleted and often
continues occupying the primary storage.

Data retention and archiving


Minimum record retention requirements are defined by laws, which vary by state and by country. For
example, the state of Massachusetts in the USA has these legal requirements.
Table 2-1 Minimum record retention requirements in the state of Massachusetts in the USA
Archiving purpose

Data retention

Business records

7 years to permanent

Contracts, leases

7 years to permanent

Employee and HR records

3 years

Payroll and benefits records

3 to 7 years

Offsite archiving should be an integral part of any backup and restore process and is a critical component
of any disaster recovery. A full backup of all important data, such as applications, operating systems, and
databases, must be stored offsite in a secure location (ideally, a geographically remote location from the
data center).
Archiving also fulfills the companys legal obligations for records retention, as listed above. The tape

media must be selected to also meet the archival life requirements:


Table 2-2 The archival life requirements
Tape technology

Archival life

9-track

12 years

8 mm

510 years

4 mm

810 years

3480 cartridge

1215 years

DLT tape

2030 years

LTO Ultrium

2030 years

What is data tiering?


There are many types of information companies have to deal with. It is important to understand the
difference between these various types of information to properly plan for a backup. Data tiering deals
with the concept of information types and arranging them into categories.
To illustrate data tiering, consider a single virtual machine running a database service. A typical database
configuration running within a virtual machine is shown in Figure 2.11.

Figure 2-11 A typical database configuration running within a virtual machine

This configuration consists of several backup candidates:


Virtual machine OS boot disk, which
o Contains the operating system image
o Changes only during the system update
o Is usually provided by the cloud service provider in the form of an ephemeral (lasting a short
time) disk image
o Does not need to be backed up if it is provided as an ephemeral image (you may want to keep a
copy which does not change when you upgrade the system image)
o Is a candidate for data tier 1
Database configuration, which
o Consists of configuration files created during the database installation and configuration
o Does need to be backed up only if the configuration changes

o Becomes assigned to data tier 2


Database installation, which
o Consists of binaries for the database installation
o Does need to be backed up only if you upgrade the database installation
o Becomes assigned to data tier 2
Database files, which
o Are located on the external storage LUN and are presented to the running virtual machine
o Contain the database records
o Must be backed up at all times
o Are assigned to data tier 3
While designing your backup solution, you can keep tier 1 data on a slow backup medium such as a nearline or offline tape, because it is unlikely that you will need to change the ephemeral image. Tier 2 data
should be placed on a faster backup medium, such as a virtual tape drive. Tier 3 data should be placed on
the fastest backup medium, such as a disk drive.

Backup BasicsYour Backup And Recovery Plan

Figure 2-12 Benjamin Franklin

Benjamin Franklin (Figure 2.12), an American writer, philosopher, scientist, politician, inventor, and
publisher, once said:
If you fail to plan, you are planning to fail!
Take this quote and see if you can apply it to your goal of protecting your companys data assets from a
disaster. It is clear that you must anticipate, define, create, implement, and validate your backup and
recovery procedures to minimize the most common consequences of data loss-related problems.
The physical process of preserving files on different media and in different locations is only a part of the
backup and recovery plan. What is most important for enterprises is to minimize or even avoid expensive
downtime. To do so, a strategy (a plan) must be created and implemented to use backup and restore as a
way to get the systems up and running again in the shortest possible time.
It is documented that 93% of companies without a backup and recovery plan go out of business within

five years of a data loss. Fifty percent of those businesses that do not recover their data within ten
business days never fully recover.
A good backup strategy must consider
The causes and cost of downtime
Answers to these questions:
o How much data can you afford to lose at any point in time?
o How much time is available for the backup? (This time is called the backup window.)
o How much downtime can you tolerate? (This information helps you define your recovery.)
A proper recovery plan must also define processes that include the following:
Partial recovery
o Single file (e.g., if a user needs to recover an older version of a file and your restore operation
should not delete the newer version)
o Folder/directory (e.g., if a project moves from one computer to another)
o All data associated with a single user (e.g., if a user moves to another department)
o All data associated with a user group (e.g., if all project files associated with a specific project
team are needed)
Full recovery
o By tape (e.g., everything stored on the backup medium/backup set)
o By a target system (e.g., everything stored on a server prior to its crash)

Figure 2-13 A sample backup and recovery plan

The results of the effort to preserve the information in your backup depend on your backup and recovery
planning (Figure 2.13). At minimum, your backup and recovery plan should include the following:

1. Identifying which data must be saved.


2. Grouping this data into (jobs.)
3. Assigning levels of importance to this data.
4. Calculating the backup frequency.
5. Protecting this data.
6. Testing your backup operations.
7. Maintaining your backup history.
8. Archiving your tapes.
9. Defining and testing your restore procedures.

Note
Depending on the situation, the order of these steps can be changed.

Step 1. Identifying which data must be saved

Figure 2-14 Steps 13

(A typical company)s data doubles, even triples, every year.


If you dont believe this statement, which may actually underestimate the data explosion, read a few
articles about global data growth and the birth of Big Data (Figure 2.14). One such article is here

Note
BIG
DATA
UNIVERSE
BEGINNING
TO
EXPLODE
(http://www.csc.com/insights/flxwd/78931-big_data_universe_beginning_to_explode).

With such data explosion, what must be backed up and preserved? All data that a business replies on
should be backed, including the operating systems, applications, and most importantly, user data.
The entire system must be restorable. Many people consider the operating system not worth being backed
up because there are operating system distribution media available to restore crashed systems. However,
the enormous work which was put into installing, configuring, and updating the system is then lost.

Steps 23. Grouping this data into (jobs) and assigning level of importance to
these job groups
In steps 23, you should decide how to best group the data by considering the level of importance, the
volume of the data, the sensitivity of the information, and the recovery requirements. This is where data
tiering, covered earlier, plays an important role.
Consider grouping the backup jobs according to

Devices
Folders/directories
Media sets
Partitions
Workgroups

Step 4. Calculating the backup frequency

Figure 2-15 Steps 45

Certain data must be backed up more frequently, compared to other data. Some factors that can help you
decide the backup frequency include the following (Figure 2.15):

Rate at which the data is created or changed in the company


Acceptable data loss (recovery point)
Value of the data
Cost and time to save the data
Allowed and acceptable downtime to restore the data

It is also necessary to evaluate the type of backup most suitable for your environment and goals. The
backup type is determined by the same factors that define the backup frequency.
Besides regular full backups of the complete system, certain user groups (e.g., developers or sales
consultants) or certain type of data (e.g., sales records, customer databases, software development files,
or financial transactions) may have the need for more frequent backups above the scheduled full system
backups. All these needs must be identified throughout the company to determine the right backup type and
frequency.
The total size of the backed up data as well as the required backup frequency define the backup
performance needed to fit the operation into the backup window. The backup window is the time
available for the backup process while the host is taken offline or the user demand upon it is the lowest.
This calculation yields a GB/hour value that, together with various other factors influencing performance,
such as the tape drives typical backup rate, the network speed and bandwidth, and the backup software
capabilities, influences the decision which backup solution must be implemented.

Step 5. Protecting this data


There is a distinction between data protection and data backup. In case of confidential data, exposing the
backup to unauthorized access is equal to the online data breach. Therefore, it is necessary to protect the
data with

Password protection (the simplest method)


Data encryption
Inventory of volumes
Physical protection, such as secure and locked storage, multiple copies of the most valuable tapes,
and offsite archiving

Since gigabytes of valuable and business-critical company data often resides on just a single tape
cartridge, it is necessary to protect this data against hacking. There are several means of doing so, one of
which is encryption to secure the tapes and make them and the enclosed corporate intelligence senseless
to possible thieves. Another is keeping track of all tapes in the company with an inventory by numbering
the media and creating tape lists with their individual locations. Last but not least, a secure and locked
storage for the cartridges with controlled access means physical security for the data on the tapes.
All of this requires that a good working backup strategy and management system are in place. Offsite
archiving can also be used to protect against theft and other breaches.

Step 6. Testing your backup operations

Figure 2-16 Steps 67

Before you close a backup job (Figure 2.16) and make the backup procedure a daily or a weekly routine,
you must make sure that the data is backed up properly. Depending on the technology used to store the
information, many things can go wrong:

Incomplete data, such as files that are open and in use by users, applications, or the operating system
Bad or worn-out media
Environmental influences, such as magnetic fields or heat
Lost passwords
Virus attacks

To prevent these problems from occurring at the restore time:


Investigate, select, and use a verification feature of the backup application
Test the backup by restoring sensitive or sample data
The way in which verification is performed differs from one backup application to another. These are
some possible ways to do it:
Reduced verificationcompares only the first n (number of) files on the tape with the original data
on the disk.
Byte-by-byte verificationcauses the backup application to read back the data from the tape and
compares it with the original data on disk for each file.
Cyclic redundancy check (CRC)causes the backup application to read back the data from the
tape, calculates the CRC value of the data, and compares that value with the CRC number stored on
the tape while writing the data.

Step 7. Maintaining your backup history

Enterprise backup solutions are designed to hold the information for the entire enterprise. This translates
to a huge volume of information in a single place. After being designed, created, and tested, the backup
jobs are then delegated to the automated systems (software or hardware tape libraries, for example).
Without knowing what data is backed up and where it is, the backed up information is almost useless.
To create and maintain a history of files or tapes, you can implement these procedures:
Suitable tape rotation schemes
Old backup retention
Media overwrite protection and tape expiration dates
A frequent and offsite backup alone is not enough to securely protect the data of an enterprise. Keeping
old backups and thus not overwriting older tapes to maintain a backup history is another key to getting lost
data back. Reasons for doing this are manifold: accidentally deleted files may be discovered weeks later,
a virus might have been found in the system after several days, or a decision was made to downgrade to a
previous version of the operating system or application after the new one turns out to be insufficient or
problematic. It saves time to be able to restore the original environment from a tape. For these and other
reasons, it makes sense to save weekly backups for periods up to several months or even years.
Backup horizon is the length of time for which you need to save your data. If you need to keep copies of
your data for three years, then your backup horizon is three years. The backup horizon is used to define the
number of tape media in rotation schemes.

Step 8. Archiving your tapes

Figure 2-17 Steps 89

Backup has a limited time validity. This backup characteristic is even embedded in the backup software
through the mechanism of data expiration and retention. Backed up information that is prepared to be kept
in a safe place for a longer time is called the archive (Figure 2.17).

Your goal is to achieve redundancyto protect against environmental disasters, theft, human errors, and
other factors. Thus, you should implement offsite backup and take your backup tapes to a different
location.
Once a regular backup procedure is in place in addition to the appropriate backup hardware and
software, it is time to consider a strategy to back up the backups to achieve some kind of redundancy.
Many events such as natural disasters (fire, water, and earthquakes) or human errors (operator puts in a
wrong tape) threaten one set of tapes. To avoid this single point of failure, a second copy of the backup
set should always be taken offsite (offsite backup). At the same time, a second backup copy, located in a
safe place in a separate location, should always be available if needed. If the offsite backup contains a
full or an image backup, it is referred to as the archive.

Step 9. Defining and testing your restore procedures


Last but not least, define and test your restore procedures. If you can, simulate a system crash and see if
your restore procedures work. If they do, time them and compare them against your initial goals and
downtime requirements. Bring into your recovery scenarios other factors that might occur (multiple
failure scenarios) and see if you can recover.
Keep in mind that your entire backup and recovery plan rests upon your ability to recover your data from
a failure.

Tape Media Drives And Technologies

Figure 2-18 Tape media drives

In the past, the magnetic tape was the best choice for backing up a computer system. Currently, tapes can
be replaced with other technologies, but they continue playing an important role in large enterprise
systems through the tape library technologies (Figure 2.18).
Tape drives are best suited for full system backups because they are sequential-access devices. If a hard
drive crashes, restoring the data from a tape drive is simple and painless.
For recovery of individual documents or files, tape drives are much less suitable, because of their
sequential-access nature and because tape backups are often done incrementally or differentially.
Recovering an individual file, therefore, involves first determining which backup and/or which tape
contains the desired file, and then forwarding the tape to the point where that file resides.

Advantages and disadvantages of tape drives


Tape drives have some advantages over other types of media, such as:
ReliabilityBecause tape drives are in use only during a backup or a recovery operation, they tend
to be more reliable compared to hard drives (which always spin, even when they are not in use).
Power savingsFor the same reason, tape drives also use less power.
VersatilityTape cartridges are usually small and can be easily stored offsite, allowing data to
survive even if the computer itself is destroyed or stolen.
Ease of useThere is a wide support for tape drives and a good selection of backup software that
makes restoring a computer from tape a reasonably painless procedure.
Tape drives also have disadvantages, such as:
ExpenseAlthough once considered as being the most economical backup method per GB of data,
tape drives and tape media are now considerably more expensive than hard drives or a network
backup.
Tape degradationMagnetic medium is subject to degradation due to heat, humidity, dust,
mishandling, electromagnetic forces, and ordinary wear.
Uncertainty of data integrityUnless a full verification of each backup is performed (which takes
as much time as the backup itself), you cannot be certain whether your backup is reliable.
Less suitability for nonfull restoresTapes are sequential-access devices and are best suited for
full system restores. Finding and restoring individual documents can be a long, slow, and
cumbersome process.
Security considerationsMost tape backups are done at night while the machine is unattended and
the previous nights tape can easily end up with an unauthorized person. As with any removable
media, security is a concern both while the machine is unattended and while the tape is being
transported.

WORM technologies

Figure 2-19 WORM technologies

The Write Once, Read Many (WORM; Figure 2.19) technology makes a nonerasable copy of data onto a
nonrewritable medium. Once written to the medium, the data cannot be modified. WORM drives

preceded the invention of CD-R and DVD-R disks. The CD-R and DVD-R optical disks for computers
are common WORM devices.
Besides the optical media, HPE offers a range of Ultrium-based WORM tape cartridges.

Removable disk backup (RDX)

Figure 2-20 Removable disk backup (RDX)

HPE RDX+ Removable Disk Backup System (Figure 2.20) is a rugged, removable, disk-based backup
solution for servers and workstations. It is designed as a backup solution for remote locations or small-tomedium businesses with a few or no IT resources that are in less-than-ideal locations such as in busy
offices, field base sites, or businesses that require mobility.
RDX requires a USB connection and no separate power source or cord. The entire system can be
protected with the hands-free HPE RDX Continuous Data Protection Software.
The HPE RDX+ Removable Disk Backup System offers fast disk-based performance with the ability to
store 500 GB, 1 TB, or 2 TB of data on a single removable disk cartridge at speeds of up to 360 GB/hr.

Tape libraries and virtual tape libraries

Figure 2-21 Tape libraries and virtual tape libraries

Enterprises deal with large volumes of information that they must back up and recover (Figure 2.21).
Traditionally, the problem of handling a large backup capacity is solved through tape libraries. Tape
libraries aggregate many tape drives and cartridges while providing the necessary mechanism to
physically move, label, and rotate tapes.
HPE offers a family of products named StoreEver (Figure 2.22), representing a wide range of tape
libraries that integrate in a backup, recovery, and archiving portfolio.

Figure 2-22 HPE StoreEver ESL G3 Tape Library

In addition to the tape-based backup devices and libraries, HPE offers a disk-to-disk backup solution
called HPE StoreOnce. These devices are configured in the backup application as Network-Attached
Storage (NAS) devices, StoreOnce Catalyst, or Virtual Tape Library (VTL) targets.

Figure 2-23 HPE StoreOnce System portfolio

The total number of backup target devices provided by a StoreOnce System (Figure 2.23) varies
according to the model. These devices may be all StoreOnce Catalyst, all VTL, all NAS, or any
combination. All StoreOnce devices automatically utilize the StoreOnce deduplication, ensuring efficient
and cost-effective use of disk space. Another benefit of StoreOnce Catalyst devices is that deduplication
can be configured to occur on the Media Server (low bandwidth) or on the StoreOnce System (high
bandwidth), allowing the user to decide which makes the most efficient use of the available bandwidth.

Learning check questions


Reinforce your knowledge and understanding of the topics just covered by completing this learning check:
1. Which term describes the length of time you need to save your data after backup?
a)Backup horizon
b)Backup window
c)Data retention interval
d)Recovery time objective
e)Recovery point objective
2. Which tape technology has the shortest archival life?
a)4 mm
b)8 mm
c)9-Track
d)3480 cartridge
e)DLT tape
3. Which technology consists of a rugged, removable, disk-based backup solution that is best suited
for regional or branch offices with no IT resources?
a)WORM
b)VTL
c)RDX
d)LTO
4. You are creating a backup and recovery plan for your customer. According to HPE
recommendations, which step should you complete after you identify which data must be saved?
a)Test your backup operations.
b)Calculate the backup frequency.
c)Define and test your restore procedures.
d)Group your data into jobs.

Learning check answers


This section contains answers to the learning check questions.

1. Which term describes the length of time you need to save your data after backup?
a)Backup horizon
b)Backup window
c)Data retention interval
d)Recovery time objective
e)Recovery point objective
2. Which tape technology has the shortest archival life?
a)4 mm
b)8 mm
c)9-Track
d)3480 cartridge
e)DLT tape
3. Which technology consists of a rugged, removable, disk-based backup solution that is best suited
for regional or branch offices with no IT resources?
a)WORM
b)VTL
c)RDX
d)LTO
4. You are creating a backup and recovery plan for your customer. According to HPE
recommendations, which step should you complete after you identify which data must be saved?
a)Test your backup operations.
b)Calculate the backup frequency.
c)Define and test your restore procedures.
d)Group your data into jobs.

Summary
This chapter focuses on backup strategies and on related topics such as restore, archiving, and data
tiering. It also covers a basic backup and recovery plan, which is something that every business ought to
develop, test, and implement. Lastly, it takes a look at tape media drives and technologies.
The key takeaways from this chapter are the following:
A backup is the process of creating a copy of your data onto a secondary medium, such as a tape or a
disk. You create a backup to prevent data loss due to possible destruction or corruption.
You back up any and all data you deem critical to your company operationoperating systems and
environments, applications, application and user data, and logs and journals.
Different backup strategies and technologies exist, each providing you with choices for defining and
implementing the right backup and recovery plan for your company.
Restoring your data means recovering your data from a backup after destruction, deletion, or

corruption. Archiving means arranging and indexing your data for long-term retention and is
different from backing up your data. Data tiering organizes your data into categories, where each
category might have a different backup and recovery plan.
Consequently, your backup solution should

Have a thoroughly tested and documented backup plan


Include a backup of your operating system, applications, and user data
Back up every hard disk in your environment
Have a thoroughly tested and refined recovery process
Back up on a daily basis, using either incremental or differential backup
Use the most appropriate tape rotation scheme
Duplicate backup tapes and store them offsite on a weekly basis
Use the data verification feature of your backup software
Employ tape libraries and/or auto-loaders to ease the administrative overhead
Plan for growth

3 Topology and Performance

CHAPTER OBJECTIVES
In this chapter, you will learn to:
List and describe various backup methods and topologies
Describe Storage Area Networks (SANs) and explain their advantages
List and explain backup and disaster recovery tiers
List and describe various HPE backup technologies and solutions
Explain and position Network-Attached Storage (NAS) backup technologies
Explain and position tape-as-NAS for archiving
Perform backup information collection needed for capacity planning

Introduction
Selecting the correct backup topology and solution is critical to your successful backup and recovery
strategy. As you would suspect, there are many options to choose from, ranging from how you perform the
backup and recovery, which infrastructure you use, to which technologies, products, and solutions you
bring into your environment. Every selection that you make impacts the effectiveness of your backup,
speed of your recovery, and cost.
This chapter provides you with background needed to make the right decisions. It covers different backup
methods and topologies and explains their advantages and disadvantages. It covers topics such as SAN,
NAS, deduplication, SAN zoning, multipathing, site replication, tape offload, and encryption. It also
discusses backup and disaster recovery tiers, which you can use in your planning.
HPE provides a number of technologies, products, and solutions designed specifically for the purposes of
protecting enterprise data and recovering from disasters. These products are overviewed in this chapter
also.
Lastly, the chapter ends with a short exercise of collecting backup information from customers in order to
perform capacity planning and recommend the customer the right backup and recovery solution.

Backup Methods And Topologies


No matter which form of backup and which tape rotation scheme is used in your backup strategy, there are
two basic methods of saving data to tape media in an enterprise networked environment:
Local backup
Remote backup (also referred to as the client/server backup)
A local backup usually corresponds to a local type of storage connectivity, usually referred to as Directly
Attached Storage (DAS). Storage (usually disk or tape) is directly attached by a cable to the computer
(server). A disk drive inside a PC or a tape drive attached to a single server are examples of simple types
of DAS. I/O requests (also called protocols or commands) access devices directly.
Local backup is the fastest way of performing backup. Speed is limited by the speed of the local bus or

the backup target. On the other hand, local backup is the least flexible solution, considering that it works
the best for the server to which the backup media is attached.
As a bridge between the local and remote backup, a technology called NAS is used. A NAS device (also
called an appliance), usually with an integrated processor and disk storage, is attached to a TCP/IPbased network (Local Area Network [LAN] or Wide Area Network [WAN]) and accessed using
specialized file access/file sharing protocols. File requests received by a NAS are translated by the
internal processor to device requests.
To perform a remote backup, it is necessary to establish a communication channel between the client and
the server. Storage resides on a dedicated networkthe Storage Area Network, or SAN. Like DAS, I/O
requests access devices directly, but the communication passes over the SAN infrastructure. Today, most
SANs use the Fibre Channel (FC) media, providing a connection for any processor and any storage on
that network. Ethernet media using an I/O protocol called Internet Small Computer Systems Interface
(iSCSI) emerged in 2001.

Local backup

Figure 3-1 Local backup

Local backup (Figure 3.1) is the traditional method of backing up your data. A backup device, usually a
tape drive, is locally (directly) attached to the computer. The backup software is also installed on that
computer and performs the backup operations by copying the data through the system from one device to
the other.
With local (basic) server backup, each server in a networked environment connects to its own backup
device using the SCSI protocol. Each server runs its own instance of a backup application, which may be
remotely controlled by a management station. However, the storage medium for each server still must be
managed locally and manually, which increases complexity and overhead in large network installations.
This approach is inefficient because each computer must have its own backup device and possibly its
own backup administrator. The speed of the backup, however, may be very high compared to other
methods because the backup data travels on dedicated local paths and does not consume LAN bandwidth.

Figure 3-2 Local backup in a networked environment

The main advantage of local backup is speed, which is only limited by local factors such as speed of the
devices and the bus/controller. Also, local backup does not consume any LAN bandwidth (Figure 3.2).
The main disadvantage is cost. Local backup is relatively expensive because each server needs its own
backup application and a backup device as well as a person who manages the media.

Local backup with tape libraries


To address user errors and the overall management of local backups, introduction of tape libraries into
the environment enables multiple computers to be directly (locally) connected to a tape library instead of
individual standalone tape devices.

Figure 3-3 Local backup with tape libraries

The environment consists of a centralized tape library with multiple tape drives, supporting a number of

servers that need to be backed up (Figure 3.3). When using automated tape libraries, there is something to
consideronly one system can control the loader robot in the tape library. What happens if the robot or
the computer controlling the robot fails?
Advantages of the local backup with tape libraries are speed, which is limited by only local factors, no
consumption of LAN bandwidth, and no need for individual tape media management.
The disadvantages consist of the computer controlling the robot or the robot itself being a single point of
failure and a high price as compared to the local backup without tape libraries.
Some of the disadvantages of physical tape libraries are addressed by virtual tape libraries (VTLs). VTL
is a network-based device which runs a software that simulates characteristics of the traditional tape
library. Instead of saving data on tapes, a VTL uses fast-rotating hard drives. Software middle layer
provides the tape library interface over the network to the backup software so that the backup software
can treat the VTL as a physical tape library.

Client/server backup
The client/server backup model is also called the centralized server backup.

Figure 3-4 Client/server backup

With this method, all computers and their local disk volumes have a LAN connection to one or more
central servers with attached backup devices. The data is pulled over the network from the clients by the
server-based backup application and then stored on tape. When a tape library is used as the backup
device, the solution is referred to as automated client/server backup (Figure 3.4). This model adds extra
capacity and advantages of automation. However, both cases present a single point of failureif the
server or the attached tape device fails, no client can be backed up.
The advantages of this solution are consolidation onto one backup device, fewer media management
challenges, and centralized management and automation.
The disadvantages are a single point of failure (either the backup server or the tape device) and LAN
performance degradation.

Push agents
Another approach is to have a backup client application (called push agent) on each machine in the
network that pushes the data to the server with the attached backup device and the server backup

application. Push agents improve the overall performance of a centralized server backup and optimize the
throughput by compressing and prepackaging the data before it is sent to the main backup application on
the backup server.

Figure 3-5 Push agent

The push model has some advantages over the pull model (Figure 3.5). With push agents, for instance,
only pure data is sent to the server, while in the pull model, the server additionally has to process all
metadata to determine which files to save (e.g., reading the metadata from the client and then checking the
archive bit).

Note
Metadata (or (data about data)) describes the content, quality, condition, and other
characteristics of data. Metadata is used to organize and maintain investments in data, to provide
information to data catalogues and clearinghouses, and to aid data transfers.
An additional advantage of the push model is that the client application does not need to know which
backup hardware is used at the server and may even have the capability to choose among several backup
servers (as part of the failover policies, for example) in the network, if available.
As for data security, each client may use its own encryption algorithm before sending the data to the
backup device (in the pull model, one common algorithm is used instead for all clients). The clients may
even have more CPU time available to compress the data to save tape capacity and reduce network
utilization.
A common example of the push agent advantages is the ability to perform deduplication (covered in the
next section) at the client side and send only deduplicated bits over the network. This approach requires
powerful clients, but greatly reduces the traffic between the client and the backup destination (such as
tape library or VTL).

Deduplication

Figure 3-6 Deduplication

Depending on the client being backed up, the push service could be enabled for deduplication.
Deduplication is the process of storing data in the (intelligent) way (Figure 3.6). To illustrate, assume
that your push agent performs a full backup procedure once in a week. Considering that a large part of the
data has not changed by this time (such as operating system and general software installation), exactly the
same information appears on the backup target repeatedly.
Another example is a backup of email servers. A typical email system could contain 100 instances of the
same 1 MB file attachment. If the email platform is backed up or archived, all 100 instances are saved,
requiring 100 MB of storage space.
With data deduplication, only one instance of the attachment is actually stored, and each subsequent
instance is just referenced to the saved copy. In this example, a 100 MB storage demand could be reduced
to only 1 MB.
Data deduplication offers other benefits. Lower storage requirements reflect on disk or tape expenses.
Deduplication optimizes disk space and allows for longer retention periods. Longer retention periods
allow for better recovery time objectives, greatly reducing the need for tape backups. Finally, data
deduplication reduces the data that is sent over a slow network (such as a WAN) and greatly improves
remote backup, replication, and disaster recovery capabilities.
Data deduplication can generally operate at the file, block, and even bit level. File deduplication
eliminates duplicate files but is not an efficient method of deduplication. Block and bit deduplication
looks within a file and saves unique iterations of each block or bit. Each chunk of data is processed using
a hash algorithm such as MD5 or SHA-1. This process generates a unique number for each piece of data,
which is then stored in an index. If a file is updated, only the changed data is saved. That is, if only a few
bytes of a document or a presentation are changed, only the changed blocks or bytes are saved (the
changes do not constitute an entirely new file). This behavior makes block and bit deduplication far more
efficient. However, block and bit deduplication takes more processing power and uses a much larger
index to track the individual pieces.
Hash collisions are a potential problem with deduplication. When a piece of data receives a hash
number, that number is then compared with the index of other existing hash numbers. If that hash number is

already in the index, the piece of data is considered a duplicate and does not need to be stored again.
Otherwise, the new hash number is added to the index, and the new data is stored. In rare cases, the hash
algorithm might produce the same hash number for two different chunks of data. When a hash collision
occurs, the system will not store the new data because it sees that its hash number already exists in the
index. This is called a false positive and can result in data loss. Some vendors combine hash algorithms
to reduce the possibility of hash collisions. They might also examine the metadata to prevent such
collisions.
For additional information on data deduplication, see these resources:
For
data
deduplication
explained
on
Wikipedia,
go
to:
http://en.wikipedia.org/wiki/Data_deduplication.
For a YouTube video on StoreOnce 2nd Generation Deduplication for Enterprise Data Protection, go
to: http://www.youtube.com/watch?v=L1RJKRnqquk.

Backup encryption

Figure 3-7 Data encryption

Backup encryption is an important aspect of data protection as a method of maintaining data security and
integrity (Figure 3.7). Depending on the complexity of the backup solution and the number of components,
different types of encryption might be necessary to protect your data.
In case of a simple point-to-point configuration, where the data drive is directly connected to the backup
server, encryption could be performed by the backup software or by the tape drive at the hardware level.
Often, encryption is combined with data compression and/or deduplication.
In case of more complex topologies, which include backup to disk and remote agents, encryption must be
performed in transit (at the SAN or LAN level) as well as on each backup target, such as temporary disks,
virtual libraries, or tape drives.
The goal of encryption is to protect the data integrity by preventing third parties to modify or access the

information without authorization.

Serverless backup
Serverless backup refers to backup operations that do not require data to be routed through a server.
Serverless backup requires a network configuration, such as a SAN, in which the storage and backup
devices are not tied to specific servers. Intelligence built into the routers or other connecting devices in
the SAN queries the servers for backup information. These devices then initiate the movement of that data
directly from the storage devices to the backup devices through the SAN. (Paraphrased from: A
Dictionary of Storage Networking Terminology; Storage Networking Industry Association)
Serverless backup provides a method of transferring data directly from the disk to the tape device with
only minimal impact on the application servers CPU and memory (typically around 1%) (Figure 3.8).
This works by having an intelligent FC device use the Extended SCSI COPY command standard, which is
included in the current SCSI-3 specification. The Extended SCSI COPY command creates a block-byblock disk image on the tape, allowing the serverless backup to perform image backups and do so with
high data throughputs of many terabytes per hour.

Figure 3-8 Serverless backup

The biggest benefit of serverless backup is reduced volume of data the application server has to process
for backup purposes, thus allowing for additional bandwidth to handle normal application traffic.
Normally, data would be read from the disk to the server and then from the server to the tape device. With
serverless backup, the application server is decoupled from the backup function and the host only needs to
send each SCSI COPY initiator command. Although this reduces the amount of data to the host, the host is
still repeatedly sending the initiating COPY commands.
Serverless backup requires a special SAN equipment, such as the HPE Network Storage Router M2402,
which supports up to eight simultaneous streams of serverless backup jobs. To implement serverless
backup, HPE Data Protector or equivalent software is necessary.

Storage Area Networks


To cope with general network-related performance issues, networks dedicated to storage devices were
introduced. A SAN is one such dedicated, high-speed network, similar to a LAN. A SAN connects storage
devices directly to servers and clients with similar interconnect technologies used in LANs and WANs,

such as routers, switches, hubs, and gateways. There are local and remote, shared and dedicated SANs,
and they include central or external storage elements. SAN interfaces can be ESCON, SCSI, HIPPI, SSA,
or FC.
SANs can also be used to increase backup performance.

Two basic FC topologies


There are two basic FC topologies:
Arbitrated loop (FC-AL)
Switched fabric

Arbitrated loop (AL)

Figure 3-9 Arbitrated loop

AL (Figure 3.9) is a FC topology in which all devices are connected in a one-way loop. Historically, FCAL was a lower-cost alternative to a switched fabric because it allows servers and storage devices to be
connected without using expensive FC switches.
In the arbitrated loop topology:
All components share the same total bandwidth.
The primary hardware connectivity components are FC hubs.
Disks and tapes cannot be connected to the same AL.

Note
For more information about the FC-AL, go to: http://en.wikipedia.org/wiki/Arbitrated_loop.

Switched fabric

Figure 3-10 Switched fabric

A switched fabric is a FC topology where devices (also called network nodes) connect to each other
using FC switches (Figure 3.10). It has the best scalability of all FC topologies but is also more
expensive due to the FC switch requirements.
A switched fabric:
Provides dedicated full bandwidth from one FC port to another.
Requires FC switches.
Allows disks and tapes to be in the same FC domain.

Note
For
more
information
about
http://en.wikipedia.org/wiki/Switched_fabric.

switched

fabric,

go

to:

Advantages of switched fabrics and SANs


Mixed primary (disk) and secondary (tape) storage devices in the same FC-AL are not supported because
disks use random access, whereas tapes use sequential access. This is very difficult to synchronize and
access to disk drives would disturb running backup jobs.
Another reason for the FC-AL device type limitation is that Microsoft Windows does not provide any
native lock mechanism schemes for shared disks between multiple servers. Meeting this goal would
require a cluster software, which is provided by the backup applications for the tape devices via the
Shared Storage option. Therefore, the backup application is where the lock management function for the
backup devices is provided.
Advantages of SANs include native support for requirements such as centralized management, scalable
capacity and performance, clustering, data sharing, multihosting, and storage consolidation. Using the
latest technologies, a FC-based SAN guarantees highest performance and provides longer connectivity
distances than traditional networks, while the storage management is handled in a similar fashion as

before, through dedicated SAN-enabled storage management tools.


FC is a high-speed serial link between devices, able to carry a number of protocols such as SCSI and IP.
FC is capable of carrying this traffic over possible distances between devices (10 km, up to 100 km) and
smaller cable diameters.
Thus, the SAN advantages include the following:

Dedicated storage network, separate from the communication network


Physical security (with optical FC)
Improved centralized management
Superior performance
Simpler cabling
Long distances

SAN terminology
Even though SANs can be configured in many different ways, such as small local clusters or large
geographically dispersed topologies, the same basic components are used:
SAN interfaces and protocolsallow connections to storage devices and enable storage to be
shared (clustered). To increase redundancy and performance, multiple channels can be installed
(loops with FC).
SAN interconnectsinclude traditional network components, such as hubs, routers, gateways, and
switches, which are used in a SAN.
SAN mediarefers to the physical medium for a SAN, which can be optical (fiber) or metallic
(copper).

Note
Fiber is the term used for optical medium, while Fibre is the term used for the FC protocol.
SAN fabricincludes, in most cases, switched FC or switched SCSI topology, which may be
extended across WANs using gateways.

Learner activity
Use the Storage Networking Industry Association (SNIA) website, accessible at the following link, to
match the terms with their definitions (Figure 3.11). The correct answers are revealed at the end of this
chapter.

Note
For a good dictionary of SAN terms, go to: http://www.snia.org/education/dictionary.

Figure 3-11 SAN terminology matching activity

Learner activity answers


This section contains answers to the learner activities (Figure 3.22).

Figure 3-22 SAN terminology matching activity answers

Learning check questions


Reinforce your knowledge and understanding of the topics just covered by completing this learning check:
1. What are the biggest disadvantages of the FC-AL topology? (Select two.)
a)Tape devices are not supported.
b)Tape and disk devices cannot coexist on the same loop.
c)Intelligent SAN switches are required.
d)Only optical cables must be used.
e)High device chatter and slow device discovery hinder performance.
2. Which statements are true about data deduplication? (Select three.)
a)Deduplication can only operate at a block level.
b)Hash collisions are a potential problem with deduplication.
c)Deduplication requires push agents.
d)Deduplication increases retention periods and improves RTO.
e)File deduplication is the most efficient method of deduplication.
f)Hash collision occurs when the same hash number is generated for different chunks of data.

Learning check answers


This section contains answers to the learning check.
1. What are the biggest disadvantages of the FC-AL topology? (Select two.)
a)Tape devices are not supported.
b)Tape and disk devices cannot coexist on the same loop.
c)Intelligent SAN switches are required.
d)Only optical cables must be used.
e)High device chatter and slow device discovery hinder performance.
2. Which statements are true about data deduplication? (Select three.)
a)Deduplication can only operate at a block level.
b)Hash collisions are a potential problem with deduplication.
c)Deduplication requires push agents.
d)Deduplication increases retention periods and improves RTO.
e)File deduplication is the most efficient method of deduplication.
f)Hash collision occurs when the same hash number is generated for different chunks of
data.

Backup And Disaster Recovery Tiers


Backup and disaster recovery (DR) are not interchangeable, but disaster recovery is not possible without
performing a backup in the first place. DR deals with the ability to get the backed-up systems restored and
running as quickly as possible.

Levels/tiers of IT disaster recovery

Figure 3-12 Disaster recovery levels/tiers

In 1992, IBM together with SHARE (IBM user group) defined seven levels (tiers) of DR solutions
(Figure 3.12) which, since then, were widely accepted in the Business Continuance industry. These seven
levels describe and quantify different options to successfully recover mission-critical computer systems
from disasters. The levels/tiers are the following:

Tier 0: Do nothing, no offsite data


Tier 1: Offsite vaulting
Tier 2: Offsite vaulting with hot site
Tier 3: Electronic vaulting
Tier 4: Electronic vaulting to hot site
Tier 5: Two-site, two-phase commit
Tier 6: Zero data loss
Tier 7: Highly automated, business integrated solution

While tier 0 is the cheapest solution, it does not incorporate any protection for your data. The higher the
level, the more expensive the solution becomes, while at the same time the recovery time is significantly
reduced. Tier 6 involves dual identical data centers with all hardware, high-speed and long-distance
connections, and copies of software, licenses, and personnel to reduce the recovery time to a few minutes.
Backup and restore are important parts of at least the first five levels of the IT DR planning.

Learner activity
Use the Internet to populate Table 3.1:

Table 3.1 DR tiers, their attributes, and expected recovery times


DR tier

Attributes (23)

Expected recovery time

0
1
2
3
4
5
6
7

Write down which source you used to populate Table 3.1:


______________________________________________

Tier 0: Do nothing, no offsite data


Tier 0 data center has no requirements for any form of DR such as backup. The site has no backup
hardware, no contingency plan, no documentation, and no backed up data. Even if a company does a
backup, but leaves the tapes in the backup device in the same computer room, it still resides in this tier.
Usually, if a disaster strikes, a company in tier 0 never recovers from the loss of their business data.

Tier 1: Offsite vaulting


A company residing in tier 1 does regular backups and moves the backup media to an offsite facility
(vault). This already is a form of DR planning, and some recovery requirements have been defined.
However, just the data (on tape) is in a safe and offsite location; however, there is no hardware such as a
compatible backup device available or a site at which to restore the data. The retrieval and vaulting of the
data are usually done by couriers and, therefore, is referred to as PTAM = Pickup Truck Access Method,
which is relatively simple and inexpensive.
Tier 1 = Offsite vaulting + recovery planning
A typical recovery time for tier 1 installations is more than a week and is dependent on when a second
site with necessary hardware can be set up. One week of business data outage can have a permanent
impact on the companys future business.

Tier 2: Offsite vaulting with hot site


Companies having a tier 2 DR plan do have a second site in standby, a so-called hot site. This secondary
site has enough compatible hardware and infrastructure to support the companys business requirements,
but is shut off. In the event of a disaster, the backups can be taken from the offsite vault to the hot site by
couriers and then be restored. A hot site generates higher cost but reduces the recovery time.
Tier 2 = Tier 1 + hot site
A typical recovery time for tier 2 installations is more than a day and is dependent on the transport time
for the data media as well as the time to start up the hot site.

Tier 3: Electronic vaulting


Companies having a tier 3 DR plan do have a second and fully operational site which is permanently up
and running (but not processing). In tier 3, electronic vaulting is supported for a subset of critical data.
This means transmitting business-critical data electronically to the secondary site where backups are
automatically created. The permanently running hot site increases the cost even more, but the time for the
recovery of business-critical data is significantly reduced.
Tier 3 = Tier 2 + electronic vaulting
A typical recovery time for tier 3 installations is around one day.

Tier 4: Electronic vaulting with active 2nd site


Tier 4 DR planning introduces two fully operational and actively processing sites which can share the
workload. Recovery of the data can be bidirectional, which means that either site can be restored from the
other. While critical data is continuously transmitted between the two sites and backups of the critical
data are taken at both sites, the noncritical data still needs to be taken from the offsite vault in case of a
disaster. In this scenario, an active management of the data stored in the hot site is required.
Tier 4 = Tier 3 + second site permanently active
A typical recovery time for tier 4 installations is up to one day.

Tier 5: Two-site, two-phase commit


In tier 5, selected data is additionally maintained in an image status. Updates to both copies of the
database in data center A (primary site) and in data center B (secondary site/hot site) are applied within a
single-commit scope. A database update request is only considered successful when the data in both sites
is successfully updated. This type of database synchronization (called remote two-phase commit) is only
possible with high-bandwidth connections between the two sites. Another requirement for tier 5 is
dedicated hardware in data center B that is able to automatically take over the workload from data center
A. All applications and critical data are present at both sites, thus only data, which is just transferred, is
lost during a disaster. Again, the recovery time is significantly reduced.
Tier 5 = Tier 4 + maintenance of selected data in an image status
A typical recovery time for tier 5 installations is less than 12 hours and could be as low as one to four
hours.

Tier 6: Zero data loss


In this one of the most expensive DR solutions, the coupling or clustering of hardware and applications is
required. Long-distance, high-bandwidth connections as well as additional hardware for data replication
adds to the cost. On the other hand, this is the fastest method of data recovery in case of a disaster. All
data is present at both sites on dual storage and is continuously updated in both directions using full
network switching capabilities. In this scenario, the applications consider data lost if a database
transaction has commenced, but the request has not been satisfied.
Tier 6 = Tier 5 + zero loss of data with automatic and immediate transfer to second site
A typical recovery time for tier 6 installations is a few minutes.

Tier 7: Highly automated, business integrated solution


Tier 7 solutions include all major components being used for a tier 6 solution with the additional
integration of automation. This allows a tier 7 solution to ensure consistency of data above that which is
granted by tier 6 solutions. Additionally, recovery of the applications is automated, allowing for
restoration of systems and applications much faster and more reliably than would be possible through
manual business continuity procedures.

Cost vs. time-to-recover relationship

Figure 3-13 Cost vs. time-to-recover relationship

The described levels (tiers) define the ability to recover data. Key factors are recovery time (how fast do
you need to recover your data) and recovery point (how much data are you willing to lose). As already
explained, a recovery solution is a compromise that you have to accept in terms of data safety and
involved cost. A good recovery solution lives on the intercept point, which presents a reasonable
proportion between the two. For planning purposes, it is vital to know the cost of downtime because you
do not want to spend more money on the solution than the financial loss involved in a disaster (usually the
longer a company cannot process data, the more expensive the outage is) (Figure 3.13).
To find out about the possible financial loss due to a disaster, a Business Impact Analysis (BIA) needs to
be undertaken and also a risk assessment should take place to identify possible problem areas.

HPE Backup Technologies And Solutions


This section covers the primary HPE backup technologies and solutions such as StoreOnce and
StoreEver. Key concepts associated with these technologies and solutions, such as SAN zoning,
heterogeneous SAN, and multipath to tape are also covered.

SAN zoning
Zoning can be helpful in larger SANs to simplify device discovery and to reduce chatter between devices.
HPE recommends the following guidelines for determining how and when to use zoning (Figure 3.14):
Use zoning by HBA port. Zoning by HBA port is implemented by creating a specific zone for each
server or host by the worldwide port name and adding only those storage elements to be utilized by
that host. Zoning by HBA port prevents a server from detecting any other devices or servers on the
SAN and simplifies the device discovery process.
Disk and tape devices on the same HBAs are supported. For larger SAN environments, it is
recommended to also add storage-centric zones for disk and backup targets. This type of zoning is
done by adding overlapping zones with disk and backup targets separated.
FC zoning can be implemented using physical switch port numbers, worldwide name IDs, or userdefined switch aliases.

Figure 3-14 SAN zoning with HPE StoreEver/StoreOnce and HPE StoreServ

Heterogeneous SAN with HPE StoreEver, StoreOnce, and StoreServ


A heterogeneous network connects different types of operating systems and different types of devices,
which all share the same SAN infrastructure. A typical usage of this network is to share tape and storage
devices among multiple servers.
HPE StoreOnce Backup System is a disk-based storage appliance for backing up networked servers or
PCs to target devices on the appliance. These devices are configured as NAS, StoreOnce Catalyst, or
VTL targets for the backup applications.

Figure 3-15 Heterogeneous SAN with HPE StoreEver, StoreOnce, and StoreServ

The SAN configuration in Figure 3.15 consists of multiple servers with different operating systems, all
connected to a shared storage. Each server can access the storage through a single host bus adapter
(HBA). HPE recommends zoning by HBA to reduce the traffic in the SAN and to improve performance.
Table 3.2 shows how the data flows within this environment. Sharing a StoreEver Tape Library or a
StoreOnce Backup eliminates the need for multiple systems connected to each server, thereby simplifying
the data protection management and reducing the total cost of ownership.
Table 3-2 Sample heterogeneous SAN solution and its data flow
Step
number

Data flows from

Data flows to

Shared SAN-attached disk array (StoreServ


storage)

Each SAN-attached backup server via FC

Each SAN-attached backup server

SAN-attached HPE StoreEver Tape Library or HPE StoreOnce Backup


via FC

Multipath to tape
Advanced Path Failover (APF) uses capabilities in the HPE StoreEver LTO-6 and newer Ultrium tape
drives and the HPE StoreEver tape libraries in which they are installed, combined with software drivers
running on a host system to provide path failover when multiple paths are available to a tape drive or to a
library controller. APF is a licensed feature.

Figure 3-16 APF with HPE StoreEver tape libraries

In Figure 3.16, two servers, designated as (backup server A) and (backup server B,) have two
different host interface ports that are connected to two different SANs. Each SAN is connected to the
StoreEver tape library. Each server has separate host bus adapter (HBA) ports dedicated to SAN 1 and
SAN 2. All drives in the library have two ports, with one port connected to SAN 1 and the other
connected to SAN 2. The library in this example has two different drives that are both configured to
provide a library control path. Each drive that is configured to provide a library control path connects to
the SAN as two devices, a tape drive and a library controller, at two different SCSI logical units.

HPE StoreOnce backup to tape offload


HPE StoreOnce backup to tape offload is a special case of a SAN configuration that leverages highspeed, disk-based backup to achieve smaller backup windows while using tape backup for economical,
long-term storage.

Figure 3-17 HPE StoreOnce backup to tape offload

Figure 3.17 illustrates a configuration that consists of one or more servers, a primary storage (HPE
StoreServ), a disk-based backup storage (HPE StoreOnce), and a shared tape library (HPE StoreEver).
This configuration allows high-speed, deduplicated backups to StoreOnce. A tape-to-tape copy can be
performed using the backup application to send rehydrated data from the StoreOnce appliance to the
SAN-attached StoreEver tape library for longer retention.
Table 3-3 HPE StoreOnce backup to tape offload and its data flow
Step number

Data flows from

Data flows to

Primary storage (HPE StoreServ)

SAN-attached server via FC

SAN-attached server

SAN-attached HPE StoreOnce backup via FC

HPE StoreOnce backup

SAN-attached server via FC

SAN-attached server

SAN-attached HPE StoreEver tape storage via FC

HPE StoreOnce Catalyst remote site replication


HPE StoreOnce Catalyst enables source- and target-side deduplication. With facilitation from the data
protection software, data can also be moved between sites without rehydration (undeduplication). In the
data center, another copy of the data can be made to tape after rehydration, utilizing tape-to-tape copy
options in your data protection software (Figure 3.18).
Typical use cases for this solution are transferring data from a branch office to the main data center or
copying data between sites.

HPE StoreOnce technology is an inline data deduplication process. It uses a hash-based chunking
technology, which analyzes incoming backup data in chunks that average 4K in size. The hashing
algorithm generates a unique hash value that identifies each chunk and points to its location in the
deduplication store.
Hash values are stored in an index that is referenced when subsequent backups are performed. When the
data generates a hash value that already exists in the index, the data is not stored a second time, but rather
a count is increased, showing how many times that hash code has been seen. Unique data generates a new
hash code and that is stored on the appliance. Typically, about 2% of every new backup is new data that
generates new hash codes.
With Virtual Tape Library and NAS shares, deduplication always occurs on the StoreOnce backup system.
With Catalyst stores, deduplication can be configured to occur on the media server (recommended) or on
the StoreOnce backup system.

Figure 3-18 HPE StoreOnce Catalyst remote site replication

Table 3-4 HPE StoreOnce Catalyst remote site replication and its data flow
Step
number Data flows from

Data flows to

Server

Server-side deduplication

Server

HPE StoreOnce backup storage at the branch site via vLAN

HPE StoreOnce backup storage at HPE StoreOnce backup storage at the data center site via WAN (movement across the WAN
the branch site
can be managed by a data protection application)

HPE StoreOnce backup storage at SAN-attached HPE StoreEver tape storage via FC (rehydrated data, utilizing the tape-to-tape
the data center site
copy operation in the data protection software)

HPE StoreOnce deduplication impact on backup performance


With VTL and NAS shares, deduplication always occurs at the StoreOnce backup system. With Catalyst
stores, deduplication may be configured to occur at the media server, which is recommended, or at the
StoreOnce backup system.
The inline nature of the deduplication process means that it is a very processor- and memory-intensive
task. HPE StoreOnce appliances have been designed with the appropriate processing power and memory
to minimize the backup performance impact of deduplication.
The best performance is achieved by configuring a larger number of libraries, shares, and Catalyst stores
with multiple backup streams to each device. However, there is always a tradeoff with the overall
deduplication ratio:
If servers with a lot of similar data must be backed up, a higher deduplication ratio can be achieved
by backing them to the same library, NAS share, or Catalyst store, even if this means directing
different media servers to the same data type device configured on the StoreOnce appliance.
If servers contain dissimilar data types, the best deduplication ratio/performance compromise is
achieved by grouping the servers with similar data types together into their own dedicated libraries,
NAS shares, or Catalyst stores. For example, if you are performing a backup of a set of Microsoft
Exchange servers, Microsoft SQL Server database servers, file servers, and application servers, it
would be best served by creating four virtual libraries, NAS shares, or Catalyst stores, with one for
each server data type.
The best backup performance to a device configured on a StoreOnce appliance is achieved using
somewhere below the maximum number of streams per device (the maximum number of streams varies
between models). When restoring data from a deduplicating device, it must reconstruct the original
undeduplicated data stream from all data chunks contained in the deduplication stores. This can result in
lower performance compared to the backup process (typically 80% of the backup performance). Restores
also typically use only a single stream.
Full backup jobs result in higher deduplication ratios and better restore performance. Incremental and
differential backups do not deduplicate well.

Nas Backup Technologies


A disk array or another storage device, which directly connects to the communication or messaging

network, using protocols such as TCP/IP, is called a NAS. A NAS functions as a server in a client/server
relationship and usually has a processor and an operating system.

Hierarchical Storage Management (HSM)


With HSM, data is stored on different media types based on the frequency with which that data is
accessed. Frequently accessed data is placed on fast devices or RAID arrays, while data with lower
demand is stored on slower devices such as optical disks or tapes. All data is always accessible to the
users, but the retrieval speed differs, depending on where the data is.
HSM is often complex to integrate and manage, but if done properly, it can reduce the cost of storage and
give users virtually unlimited storage space. HSM is not a replacement for backup; it is a storage
management system. HSM systems must be regularly backed up as well.

Near-online storage
With near online storage, high-speed and high-capacity tape drives or tape libraries are used as disk
replacements. Data that is not important or too large for disks is stored on tapes in an autoloader system
or in a library. The application software must be able to read and write data from or to the tape as if it
were a normal disk drive (the software is not a backup application). The tapes are not removed for
archiving, but left in the devices for user access. In this scenario, which usually is used with desktop
systems and infrequently in client/server environments, the tape drive acts as a slow disk.

Learning check questions


Reinforce your knowledge and understanding of the topics just covered by completing this learning check:
1. What benefit does APF provide?
a)Deduplication at the data source
b)Deduplication at the target
c)Multipath failover for tape devices
d)Direct application access to tapes
2. Which technology uses high-speed and high-capacity tape drives or tape libraries as replacement
for disk drives?
a)Near online storage
b)SAN
c)Deduplication
d)Backup to tape offload
3. Which DR tier requires confirmation of data updates at two different sites as part of a single
transaction and yields a recovery time of 12 hours or less?
a)Tier 3: Electronic vaulting
b)Tier 4: Electronic vaulting to hot site

c)Tier 5: Two-site, two-phase commit


d)Tier 6: Zero data loss
4. Which DR tiers bring the recovery time down to one hour or less? (Select all that apply.)
a)Tier 1: Offsite vaulting
b)Tier 2: Offsite vaulting with hot site
c)Tier 3: Electronic vaulting
d)Tier 4: Electronic vaulting to hot site
e)Tier 5: Two-site, two-phase commit
f)Tier 6: Zero data loss
g)Tier 7: Highly automated, business integrated solution

Learning check answers


This section contains answers to the learning check questions.
1. What benefit does APF provide?
a)Deduplication at the data source
b)Deduplication at the target
c)Multipath failover for tape devices
d)Direct application access to tapes
2. Which technology uses high-speed and high-capacity tape drives or tape libraries as replacement
for disk drives?
a)Near online storage
b)SAN
c)Deduplication
d)Backup to tape offload
3. Which DR tier requires confirmation of data updates at two different sites as part of a single
transaction and yields a recovery time of 12 hours or less?
a)Tier 3: Electronic vaulting
b)Tier 4: Electronic vaulting to hot site
c)Tier 5: Two-site, two-phase commit
d)Tier 6: Zero data loss
4. Which DR tiers bring the recovery time down to one hour or less? (Select all that apply.)
a)Tier 1: Offsite vaulting
b)Tier 2: Offsite vaulting with hot site
c)Tier 3: Electronic vaulting
d)Tier 4: Electronic vaulting to hot site
e)Tier 5: Two-site, two-phase commit

f)Tier 6: Zero data loss


g)Tier 7: Highly automated, business integrated solution

Tape-As-Nas For Archiving


One of the biggest challenges with archiving that customers experience is the management of technology
updates. This technology update management involves rolling archives from older technologies and
media forward to newer technologies and media and doing so before the older technologies become
obsolete. Tape-as-NAS (tNAS) simplifies these technology updates and ensures long-term data
accessibility.

Challenges with using backup applications for archiving


If you already have a tape-based archive or are using your backup application to archive your data, you
may already know that moving from older generation tape drives to newer ones requires an application
intervention:
You need to identify the data that must be migrated.
You must read the data from an old tape and write it to a new tape.
If keeping up with technology updates is not addressed, long-term data management becomes a problem
over time as the data continues to grow. If you are using the backup application for archiving, you must
keep that application available in case that you need to recall the old data.
While tape is a great choice for archiving, using a backup application for archiving is not always the best
answer. The data is stored in a proprietary format and requires both an administrator and a compatible
technology to access it.
Using a tNAS solution for archiving simplifies access to the archived data and the future technology
updates that are needed for data preservation.

Using tNAS for archiving

Figure 3-19 LTFS volume shown in Windows Explorer

A tNAS solution is accessed via Common Internet File System (CIFS) or Network File System (NFS)
mount points, making the data easily accessible using normal file system methods. Such access allows
future users to find their information based on user-defined file names and directory structures. A backup
administrator or a backup application is no longer needed to retrieve the data. The data is stored on tape
in the industry-standard Linear Tape File System (LTFS) format, thereby eliminating the need for any
proprietary software for future access (Figure 3.19).

Moving to a tNAS solution

Figure 3-20 tNAS solution based on QStar ASM

Moving to a tNAS solution is simple. All you need to do is make your tNAS mount point the destination
for the data as it is read or recovered from the older tape. As the data is read from the older tapes, it is
sent to the tNAS cache behind the mount point (QStar ASM cache in Figure 3.20) and then moved
automatically to the new tape drive located behind the cache in the LTFS format (HPE StoreEver tape
library with new technology drives in Figure 3.20).
The tNAS solution from HPE uses standard hardware with the Archive Storage Manager (ASM)
software from QStar. It supports LTFS and all HPE tape libraries.

Technology updates with tNAS


When it is time to move your data to a new technology, it is a simple process also, and it is managed
within the tNAS solution. For example, if you have data on LTO-5 drives and want to install LTO-7
drives when they become available, you can use a feature called TMT to automate this data migration.
TMT is part of the tNAS solution stack, and it optimizes the data transfer between the drives as a
background task.

Tiered storage and data lifecycle management with tNAS

Figure 3-21 Tiered storage and data lifecycle management with tNAS

QStar Network Migrator (QNM) is a policy-based tiered storage and data lifecycle manager. QNM
software uses advanced policy management to monitor and automatically migrate, copy, and move
infrequently used files from a primary storage to a tiered storage or to a central archive (Figure 3.21).
By migrating static or less frequently used files to lower-cost storage such tape, businesses can optimize
the use of primary storage, reducing the need and costs associated with purchasing more. In addition,
when data is managed properly and only the most recent or frequently changing files are included in the
backup process, the backup window can be reduced.

tNAS with LTO tape for archiving


HPE StoreEver LTO tape is a great choice for long-term data storage. It is reliable, durable, and costeffective. Its ability to read back two generations of media makes LTO an attractive option for archiving.
With the release of LTO-7, you are able to store up to 15 TB* of data on a single tape cartridge and
access it at up to 750 MB/s* speed, achieving over 18 times faster access than LTO-1.
tNAS is a great solution built on the value of LTO and makes ongoing data management and access easier.
You still need to plan your migrations and deletions of your old data.

Key benefits of tNAS


The key benefits of tNAS include the following:

Traditional NAS mount point accessibility


Easy integration into existing workflows
No administration that is required for data reads from and writes to tape
No manual media management

Self-allocation of media as needed


User-created directory structure for custom data management
Integration with HPE Data Verification for ongoing media health analysis
Automatic creation of secondary copies for offsite protection
Monitoring of data life and recycling of tape cartridges

Sample Backup Information Collection For Capacity Planning


This section guides you through a sample exercise of collecting backup information for capacity planning
purposes.
Write your thoughts, answers, and notes in the provided spaces.

Scenario
Your customer asked you to prepare a proposal for a backup solution. You are analyzing the backup
performance requirements and planning for the right capacity to meet this customers needs.
You wrote the following questions in preparation for your meeting with the customer:
How long does it take to back up the customers environment? In other words, what is the backup
window?
How much data must be backed up?
How many tapes are needed?
How many tape drives are necessary?
What other questions must you ask?
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
What information must you collect to answer these questions?
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________

_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________

Information needed per node


To do your capacity planning, you need to collect the following type of information:
Bus/adapter(s):
o Maximum bandwidth (MB/s)
o Maximum throughput (IO/s)
o Available bandwidth
o Available throughput
Storage (per logical drive):
o Total capacity (in GBs)
o Used capacity (for full backups)
o % daily change (for incremental backups)
o % overlap (for differential backups)
o Type of data (for compression calculations)
o Maximum data transfer rate
o Maximum throughput
o Available bandwidth
o Available throughput
Network connection:
o Maximum bandwidth
o Available bandwidth
o % protocol overhead

Note
Maximum is used to indicate the upper limit of a given device, not the theoretical limits of the
technology used (e.g., FC-AL has a theoretical limit of approximately 50,000 IO/s; however, a
given FC-AL adapter may be capable of less than 50% of this rate).
Available is used to indicate the resource capability of a given device not currently in use.
An adapter is a physical device that allows one hardware or electronic interface to be adapted
(accommodated without a loss of function) to another hardware or electronic interface. The card
adapts information that is exchanged between the computers microprocessor and the devices
that the card supports.

A controller is a device that controls the transfer of data from a computer to a peripheral device
and vice versa. For example, disk drives, display screens, keyboards, and printers all require
controllers.
Is there any other type of information that you believe you need to collect? If so, indicate it below:
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________

Information needed per network device


Information you need to collect for network hubs and switches includes the following:

Maximum transfer rate


Maximum throughput
Available bandwidth
Available throughput
Number of ports; then, for each port:
o Maximum transfer rate
o Maximum throughput
o Available bandwidth
o Available throughput

Example
10 ports at 100 MHz each yield 125 MB/s.
What if the switch throughput is only 50 MB/s?
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
Is there any other type of information that you believe you need to collect? If so, indicate it below:
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________

_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________

Information needed per backup server


For each backup job, you need to collect answers to the following questions:

Which clients are being backed up?


Which files are being backed up?
What is the data compression rate?
Which rotation scheme should be used? What is the frequency, protection period, and append rules?
Which backup type (full, differential, or incremental) should be used and when?

Miscellaneous questions may include the following:


What is the expected data growth rate?
What is the data change rate?
Is there any other type of information that you believe you need to collect? If so, indicate it below:
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________
_______________________________________________________________________________

Capacity planning calculations


As part of the capacity planning calculations, you determine

The data arrival rate at the backup server(s)


The number of tape drives needed
The number of tape media needed
The duration of the backup window

At the same time, you may be able to identify possible bottlenecks in the device chain.
If necessary, try different variables and recalculate for different approaches or solutions.

Summary
The basic backup methods consist of performing local backup and remote backup.
Local backup uses backup devices directly attached to the host you want to back up. It is typically not
only the fastest method of backing up data but also the least flexible option and often expensive. Local
backup can use advanced technologies such as tape libraries to automate management and share the cost
among more devices.
Remote backup uses shared devices, such as tape libraries, autoloaders, and VTLs, and copies the data
over a network. Deduplication plays an important role with remote backups, as it lowers storage
requirements, optimizes disk space, and allows for longer disk retention periods and better recovery time
objectives.
SAN is a high-speed network that connects storage devices directly to servers and clients. SAN is used
for storage access and for remote backup purposes.
SAN topologies include FC-AL and switched fabric. Switched fabric is a better topology because it
supports multiple device types, provides dedicated full bandwidth, and supports requirements such as
centralized management, scalable capacity and performance, clustering, data sharing, multihosting, and
storage consolidation.
Advantages of FC and SANs are the following:

Dedicated storage network, separate from the communication network


Physical security (with optical FC)
Improved centralized management
Superior performance
Simpler cabling
Long distances

Different technologies also exist to further optimize or complement backup and restore. These
technologies include serverless backup, multipath to tape, backup to tape offload, remote site replication,
SAN zoning, HSM, and near online storage.
DR deals with the ability of data centers to become fully operational after a disaster. Levels or tiers of
DR are from 0 (the lowest) to 7 (the highest). As you move up each level, your ability to recover from a
disaster improves, but at a higher expense. Therefore, before deciding upon the right DR tier, a customer
should perform a cost vs. time-to-recover analysis and a business impact analysis.
Backup encryption maintains data security and integrity by protecting the data against unauthorized access
and alterations.
tNAS is a solution that provides an easy way to access tape as if it were a NAS storage. You mount the
tNAS volume as a traditional NAS mount point, which gives you simple drag-n-drop access to your files.
Thus, tNAS simplifies access to archived data and helps companies move to future tape technologies.

4 Backup and Archive Configurations

CHAPTER OBJECTIVES
In this chapter, you will learn to:
List, describe, and position the following configurations and solutions:
o Simple backup, such as point-to-point, HPE StoreEver, and HPE StoreOpen
o SAN-based backup, such as connected independent fabrics, HPE StoreEver tape library
partitioning, and extended SAN
o Remote backup, such as HPE StoreServ Remote Copy software, HPE StoreOnce Catalyst remote
site replication, and HPE StoreServ File Persona
o Virtualization-related backup, such as HPE StoreOnce Recovery Manager Central for VMware,
Veeam backup and restore, and VMware vStorage API for Data Protection (VADP)
Describe the HPE Enterprise Secure Key Manager
Explain how tiered data retention works with HPE StoreServ
Demonstrate how to effectively use the HPE Data Agile BURA Compatibility Matrix

Introduction
Familiarity with basic concepts is essential, but this is not sufficient for effective design of BURA
solutions. This chapter provides you with sample configurations of HPE backup and archiving solutions
for you to understand which components are used for each configuration, what their function is, and how
they are connected. Some of the more complex designs are assembled from many components; this chapter
provides an enumerated list of steps to illustrate the data flow within these designs.
The examples covered in this chapter include direct-attached basic and extended SAN designs. All
hardware items shown in these examples are listed in the HPE Data Agile BURA Compatibility Matrix.
This compatibility matrix provides information for designing data protection solutions, including backup,
restore, and archiving using HPE StoreOnce (disk), HPE StoreEver (tape), and HPE StoreAll (NAS)
storage products.

Note
For
the
HPE
Data
Agile
BURA
Compatibility
Matrix,
go
http://h20566.www2.hpe.com/hpsc/doc/public/display?
sp4ts.oid=412183&$3;mp;amp;docId=emr_na-c04616269&$3;mp;amp;docLocale=en_US.

to:

Simple Backup Configurations


Simple backup configurations include point-to-point solutions and direct copy of files to the tape media
using the Linear Tape File System (LTFS) software.

Point-to-point configuration

Typical usage:
Local and network client backup
The point-to-point configuration is the simplest method of connecting the backup server with the backup
device. Although the configuration is simple, it can provide enterprise-grade performance and capacity.
Examples shown in Figure 4.1 use these HPE technologies:
HPE StoreEveris enterprise-grade tape library system, built for capacity, performance, and
reliability. HPE StoreEver can scale up to over 12,000 tape cartridges in increments of 100 slots for
capacity on demand. Incorporating between 1 and 192 drives, you can consolidate and store up to
75 PB (with data compression of 2.5:1) of enterprise data.
HPE StoreOpenHPE StoreOpen with LTFS for Windows is a tape-based file system for HPE
LTO tapes. HPE StoreOpen Standalone for Windows is a free application that helps LTFS users use
and manage individual HPE tape drives. Available for Mac and Windows environments, it
completely removes the need for low-level terminal commands by guiding you through the full
process of selecting, preparing, and mounting LTFS cartridges. It essentially transforms into a
device which looks and works like a regular block device, making it easy to search and work with
backed up files.
In Figure 4.1, the backup servers transfer data to the HPE StoreEver tape drive or the HPE StoreEver tape
library. The servers utilize a backup software or HPE StoreOpen with LTFS for managing LTFS
functionality on the HPE StoreEver LTO-5 and newer Ultrium tape drives. External tape drives use a SAS
interface. Drives contained in the HPE StoreEver tape library use a Fibre Channel (FC) interface.
Clients on the LAN transfer data first to the backup server and then send the data to the tape storage. This
solution provides data protection for a large number of clients through a dedicated backup server.

Figure 4-1 Point-to-point connectivity and data flow with HPE StoreEver

Table 4.1 explains how the data flows within this point-to-point solution.
Table 4-1 Point-to-point solution with HPE StoreEver data flow
Step number

Data flows from

Data flows to

Backup server

HPE StoreEver tape storage device via FC or SAS

HPE StoreOpen with LTFS software

Typical usage:
Data transfer between a workstation/host and a standalone HPE StoreEver tape drive or HPE
StoreEver tape library for backup or data transport between LTFS solutions.
There are two ways of accessing information on the backup media:
Sequential
Random
In case of sequential access, each read operation requires that you start reading from the beginning,
regardless of where your data is. In case of a tape drive, that means that you have to rewind the tape or
keep the record of the current position while handling the tape. Sequential access is suitable for legacy

backup technologies such as the UNIX tar command, but it is not user-friendly.
The random access mode allows you to access the information directly without the need for sequential
reading from the beginning of the media. This is the common access method for optical (DVD) and
magnetic (HDD) devices.
To bridge the gap between sequential and random access to tapes, HPE offers LTFS and StoreOpen:
HPE StoreOpen is a set of free software utilities that provides disk-like access to HPE StoreEver
tape products using LTFS.
HPE StoreOpen Standalone mounts a tape drive to the host like a disk or a thumb drive, thereby
giving you access to the data on the tape cartridge just like a disk drive or a thumb drive.
HPE StoreOpen Automation mounts a tape library to a host and presents each tape cartridge that is in
the library as a file folder behind a mount letter. Each file folder is given the name of the barcode
label of the tape media and can be accessed like any other folder in the operating system. The user
can navigate the directory of an LTO cartridge even if it is not loaded in an LTO tape drive. A
cartridge is loaded automatically if a data read or write operation is needed. HPE StoreOpen reads
and writes in the LTFS format, thereby making it a great complementary product to other tape
solutions that use LTFS, such as the HPE Tape-as-NAS solution, allowing for data exchange
between solutions.
Figure 4-2 displays the StoreOpen Standalone operating system view and Figure 4.3 displays the
StoreOpen Automation operating system view.

Figure 4-2 StoreOpen Standalone operating system view of an LTFS volume

Figure 4-3 StoreOpen Automation operating system view of an LTFS volume

Figure 4-4 StoreOpen with LTFS connectivity and data flow

Table 4.2 explains the StoreOpen Standalone data flow.

Table 4-2 StoreOpen Standalone data flow


Step number

Data flows from

Data flows to

Host/workstation

HPE StoreEver Standalone LTO tape drive via directly connected SAS

Table 4.3 explains the StoreOpen Automation data flow.


Table 4-3 StoreOpen Automation data flow
Step number

Data flows from

Data flows to

Host/workstation

HPE StoreEver tape library via directly connected SAS or via FC

San-Based Backup Configurations


SAN-based backup configurations typically utilize connected independent fabrics, tape library
partitioning, and/or extended SAN. These are explained in this section.

Connected independent fabrics

Typical usage:
Connecting multiple SANs.
Figure 4.5 illustrates a configuration that consists of independent fabrics, or SAN islands, utilizing either a
multiprotocol router (MPR) or Fibre Channel routing (FCR) between switches. Multiple backup servers
connect to their local switch within their independent fabric. Through MPR or via licensing for FC
routing, each server can connect to the shared tape library as if it were in the same independent fabric.
This solution enables the interconnection of devices between SAN fabrics without merging those fabrics,
thereby providing a more secure, flexible storage networking foundation.

Figure 4-5 Connected independent fabrics

HPE StoreEver tape library partitioning

Typical usage:
Segregation of different generations of LTO media, or when multiple departments or data
protection applications need to share a single tape library.
Reliance on simple tape devices is not an option for the enterprises. Instead, they use tape libraries.
StoreEver is the common name for different tape library devices built by HPE. Tape libraries have a
large capacity and can contain multiple tape drives to allow parallel access. To provide parallel access
to different clients at the same time, tape libraries can be partitioned.
Figure 4.6 illustrates a configuration consisting of two independent backup applications or departments
accessing a single tape library. Utilizing the partitioning feature in the tape library management console
and HPE Command View for Tape Libraries, each backup application or department is presented with a
logical library comprised of a subset of drives and slots from the physical library. HPE Command View
for Tape Libraries software is a single-pane-of-glass management software which manages, monitors, and
configures all HPE tape libraries through a single console. Tape library partitioning increases flexibility

in the data center and lowers the total cost of ownership (TCO).

Figure 4-6 HPE StoreEver tape library partitioning

Extended SAN

Typical usage:
Offsite backup and disaster recovery.
Similar to LANs and WANs, SANs also span over long distances. Extended SAN leverages transport
technologies and design solutions to achieve the long distance, high-performance communication over FC.
Figure 4.7 illustrates a shared HPE StoreEver tape library and HPE StoreOnce Virtual Tape Library
(VTL) that use SAN extension technologies to facilitate connectivity between remote sites. Data is read
from the disk storage array located at site A and written to the StoreEver tape library or the StoreOnce
VTL located at site B over the extended network link. While Fibre Channel over IP (FCIP) and Wave
Division Multiplexing (WDM) allow connectivity over very long distances, long-wave small formfactor pluggable (SFP) transceivers can accommodate distances of 1035 km. This solution illustrates a
variety of SAN extension technologies, providing offsite connectivity for remote backup and disaster
recovery (DR).

Figure 4-7 Extended SAN connectivity and data flow

Table 4.4 explains the data flow within the extended SAN.
Table 4-4 Extended SAN data flow
Step number

Data flows from

Data flows to

Each disk array

Each SAN-attached server via FC

Each SAN-attached server

SAN-attached HPE StoreEver tape library or HPE StoreOnce VTL via FC

Learning check questions


Reinforce your knowledge and understanding of the topics just covered by completing this learning check:
1. In which tape cartridge slot increments must you scale up the HPE StoreEver tape library system?
a)50
b)100
c)150
d)200
2. Which HPE software enables partitioning of StoreEver tape libraries?

a)StoreOpen Enterprise
b)StoreOpen Automation
c)Command View for Tape Libraries
d)Recovery Manager Central for Virtual Tape Libraries

Learning check answers


This section contains answers to the learning check questions.
1. In which tape cartridge slot increments must you scale up the HPE StoreEver tape library system?
a)50
b)100
c)150
d)200
2. Which HPE software enables partitioning of StoreEver tape libraries?
a)StoreOpen Enterprise
b)StoreOpen Automation
c)Command View for Tape Libraries
d)Recovery Manager Central for Virtual Tape Libraries

Remote Backup Configurations


Remote backup configurations discussed in this section include the following:

HPE StoreServ Remote Copy software


HPE StoreOnce Catalyst remote site replication
HPE StoreAll backup to HPE StoreEver tape storage with Symantec NetBackup
HPE StoreServ File Persona Share Backup
HPE StoreServ File Persona software NDMP backup
HPE StoreOnce VSA and enterprise remote office/branch office (ROBO)

HPE StoreServ Remote Copy software

Typical usage:
Disaster tolerance, business continuity, and long-distance replication.
The HPE StoreServ Remote Copy software enables multisite and multimode replication with both
midrange and high-end arrays. Remote-copy configurations are based on the relationship between a pair

of storage systems, known as the remote-copy pair. Within a remote-copy pair, the primary storage system
A is the system holding the volumes that are copied onto the secondary storage system B.
Figure 4.8 illustrates an asynchronous remote copy operation between arrays at different sites with a tape
backup at site B. By creating a snapshot of data at site B and presenting this data to a server at site B, this
data can be backed up to a library at site B, thereby creating a remote backup of data from site A.

Figure 4-8 HPE StoreServ Remote Copy software connectivity and data flow

Table 4.5 explains the HPE StoreServ Remote Copy data flow.
Table 4.5 HPE StoreServ Remote Copy data flow
Step number

Data flows from

Data flows to

HPE StoreServ storage at site A

HPE StoreServ storage at site B via a WAN

Snapshot of replicated LUNs

Takes place within the array at site B

HPE StoreServ storage at site B

SAN-attached DR server(s) via FC

DR server(s)

SAN-attached HPE StoreEver tape library via FC

HPE StoreOnce Catalyst remote site replication

Typical usage:
Transferring data from a branch office to the main data center or copying data between sites.

While planning for the backup, archiving, and DR strategies, companies have to take care of backup
media handling, physical storage, and security. Traditionally, copies of the tapes have been physically
moved between the primary and secondary sites. However, this is slow, expensive, and insecure.
To address the speed, security, and cost issues, companies started to leverage LAN and WAN
infrastructures. It is easy and relatively cheap to build a LAN infrastructure which can move your backup
information over the network. As the volume of data grows and the number of backup locations
multiplies, it becomes expensive to maintain the links between the remote sites.
HPE offers StoreOnce backup system to address performance, security, and complexity of the backup and
archival process. The HPE StoreOnce backup system is a disk-based storage appliance for backing up
network media servers or PCs to target devices on the appliance. These devices are configured as either
NAS shares, Catalyst stores, or VTL targets for the backup applications.
HPE StoreOnce Catalyst allows source- and target-side deduplication. With facilitation from the data
protection software, data can also be moved between sites without rehydration. In the data center, another
copy of the data can be made to tape after rehydration, utilizing the tape-to-tape copy options in your data
protection software.

Note
HPE StoreOnce Catalyst remote site replication was covered in more detail in Chapter 3.

HPE StoreAll backup to HPE StoreEver tape storage with Symantec NetBackup

Typical usage:
HPE StoreAll Gateway NAS data protection.
Figure 4.9 illustrates a Network Data Management Protocol (NDMP) configuration for a three-way
NDMP backup of the StoreAll Gateway to a StoreEver tape library. NDMP is an open protocol used to
control data backup and recovery communications between the primary and secondary storage in a
heterogeneous network environment.
The configuration consists of two StoreAll Gateway clusters with HPE StoreServ storage attached, a
media or master server, and a StoreEver tape library. Cluster 1 acts as the NDMP Host Data Mover or
Data Service Provider (DSP). The NetBackup for NDMP master or media server acts as the data
management application (DMA) server and manages the robotics. During the backup session, StoreAll
Gateway cluster 1 acts as the Tape Data Server and StoreAll Gateway cluster 2 acts as the NDMP Data
Server.

Figure 4-9 HPE StoreAll backup to HPE StoreEver Tape Library with Symantec NetBackup connectivity and data flow

Table 4.6 explains the data flow within this environment.


Table 4-6 StoreAll backup to StoreEver tape storage with Symantec NetBackup data flow
Step number

Data flows from

Data flows to

HPE StoreAll Gateway cluster 2 (via NDMP)

HPE StoreAll Gateway cluster 1 (via LAN)

SAN-attached HPE StoreAll Gateway cluster 1

SAN-attached HPE StoreEver tape library via FC

Figure 4.10 illustrates the data flow for a standard tape backup scenario for the HPE StoreAll Gateway,
where the backup software is installed directly on the Gateway servers. The StoreAll cluster and
StoreEver tape library are both SAN-attached.

Figure 4-10 Standard HPE StoreAll backup to HPE StoreEver Tape Library with Symantec NetBackup connectivity and data flow

Table 4.7 explains the data flow for a standard tape backup for StoreAll Gateway when the backup
software is installed on the Gateway servers.
Table 4-7 A standard tape backup data flow for StoreAll Gateway with backup software on Gateway
servers
Step number

Data flows from

Data flows to

SAN-attached Gateway cluster

SAN-attached StoreEver tape library via FC

HPE StoreServ File Persona Share Backup

Typical usage:
NAS file storage.
Figure 4.11 shows a Share Backup with the HPE StoreServ File Persona software. The HPE StoreServ
File Persona software suite is a licensed feature of the HPE StoreServ operating system and enables a

rich set of file protocols and core file data services for client workloads. File Persona bridges the gap
between the native block storage StoreServ capacity and common file system-based storage.
Figure 4.11 illustrates an Ethernet LAN for reading and transferring data from HPE StoreServ 7400c,
running the HPE StoreServ File Persona services, to the media server. The media server then uses the
appropriate protocol to read the data from the File Persona and writes it to the target. The target can be an
HPE StoreEver tape library using FC, an HPE StoreOnce appliance using FC/SMB or NFS (depending on
the StoreOnce model), or an HPE StoreAll appliance using the SMB or NFS protocol.

Figure 4-11 HPE StoreServ File Persona Share Backup connectivity and data flow

Note
You can use the entire HPE StoreServ converged controller family in place of HPE StoreServ
7400c.
Table 4.8 explains the data flow within the HPE StoreServ File Persona Share Backup solution.
Table 4-8 HPE StoreServ File Persona Share Backup data flow
Step
number Data flows from

Data flows to

HPE StoreServ 7400c running File


Persona (Ethernet)

Media server via a LAN

Media server

SAN-attached HPE StoreEver tape library via FC, HPE StoreOnce storage via FC/LAN,
or HPE StoreAll storage via LAN

HPE StoreServ File Persona software NDMP backup

Typical usage:
NAS file storage.
Figure 4.12 illustrates a NDMP backup between the File Persona and the HPE StoreOnce/iSCSI library
via the Gigabit Ethernet NIC. The NDMP protocol transports the data between the NAS devices and the
backup devices. This removes the need for transporting the data through the backup server itself, thus
enhancing speed and removing load from the backup server.
Figure 4.12 shows the HPE StoreServ 7400c solution running the HPE StoreServ File Persona services
acting as the NDMP data mover and the ISV NDMP agent server acting as the data management
application (DMA). The DMA controls and maintains the overall backup schedule for the File Persona
systems and arranges the activities between them and the iSCSI library when the backup session starts.
Then, the DMA is released from the session and once the session is completed, the metadata is sent to the
DMA for review. The independent software vendor (ISV) NDMP agent controls the metadata and the
NDMP flow control to the NDMP server, where the data can then be sent to the iSCSI library for backup.

Figure 4-12 HPE StoreServ File Persona NDMP backup connectivity and data flow

The ISV used for the NDMP backup must support NDMP. You can use the entire HPE StoreServ
converged controller family in place of the HPE StoreServ 7400c solution.
Table 4.9 explains the data flow for the HPE StoreServ File Persona software NDMP backup.

Table 4-9 Data flow for the HPE StoreServ File Persona software NDMP backup
Step number

Data flows from

Data flows to

HPE StoreServ 7400c running HPE StoreServ File Persona services (Gigabit Ethernet)

iSCSI library via LAN

HPE StoreOnce VSA and enterprise remote office/branch office (ROBO)

Typical usage:
Managing local backup and offsite backup copies.
HPE StoreOnce VSA is a virtual appliance running the HPE StoreOnce software. HPE StoreOnce VSA is
optimal for centrally managed enterprise remote office or branch office (ROBO) locations with local
backup and offsite backup copies. Branch offices can send backup data to a local HPE StoreOnce VSA
target with data deduplication-optimized copy over a WAN or to a remote HPE StoreOnce backup
appliance located at the data center and DR site.
Figure 4.13 illustrates the HPE StoreOnce VSA and enterprise ROBO configuration and its data flow.

Figure 4-13 HPE StoreOnce VSA and centrally managed ROBO connectivity and data flow

Virtualization-Related Backup Configurations


These days, virtualization is widely accepted in the enterprise environments. Server-based virtualization
enables consolidation of multiple physical servers onto fewer ones running virtual machines (VMs) to
reduce cost and increase utilization.
Virtualization-related backup configurations discussed in this section include the following:

HPE StoreOnce Recovery Manager Central for VMware


VMware vStorage API for Data Protection (VADP)
Veeam backup and restore to HPE StoreEver tape storage with HPE StoreServ
Veeam backup to HPE StoreOnce with tape offload

HPE StoreOnce Recovery Manager Central for VMware

Typical usage:
Protect VMware VM disks and datastores.
HPE StoreOnce Recovery Manager Central (RMC) provides native integration of HPE StoreServ and
HPE StoreOnce in a converged data protection solution that removes the dependency on traditional
backup software.
Figure 4.14 illustrates how RMC for VMware (RMC-V) enables protection of VMware VM disks and
datastores using consistent snapshots (which can then be used for rapid online recovery) and StoreOnce
Express Protect Backup by facilitating backups of snapshots from StoreServ to StoreOnce.
Backups on StoreOnce are self-contained volumes that are deduplicated to save space. These backups can
be used to recover data back to the original StoreServ storage system or to a different system.
Administrators can access all functionality from within the VMware vCenter web client.

Figure 4-14 HPE StoreOnce Recovery Manager Central for VMware connectivity and data flow

Table 4.10 explains the data flow within the HPE StoreOnce RMC-V solution.
Table 4-10 HPE StoreOnce RMC-V data flow
Step
number

Data flows from

Data flows to

The vCenter administrator initiates StoreOnce Express Protect Backup of a VM


or a datastore.

The VM I/O is quiesced, and the VM snapshot is taken.

The RMC virtual appliance instructs the StoreServ array to take a snapshot of
the virtual volume.

VM snapshots are deleted and the VM I/O resumes.

The StoreServ array copies the virtual volume snapshot (VM data)

to the RMC virtual appliance via FC.

RMC virtual appliance moves data via the Catalyst client low-bandwidth
transfer

to the StoreOnce appliance via FC.

The snapshot is exported to the RMC virtual


appliance.

VMware vStorage API for Data Protection (VADP)

Typical usage:
SAN backup of VMware.
VMware provides a layer of communication between the VM and the underlying storage infrastructure. It
works as a form of a file system, called the VMFS, and provides advanced features such as clustering
capabilities and snapshot functionality.
To leverage the underlying hardware capabilities, VMware provides application programming

interfaces (APIs) to third-party vendors. The VMware API for Data Protection is called VADP. VADP
enables the backup software to perform centralized VM backups without the overhead of running
individual backups from inside each VM.
Figure 4.15 illustrates a simple VADP backup environment with a vCenter Server, an ESX server running
multiple VMs, a backup server with data protection software that is integrated with VADP, a StoreServ
array where the datastore LUNs reside, and a StoreOnce storage or a StoreEver tape library for backup.

Figure 4-15 VMware VADP data protection connectivity and data flow

Table 4.11 explains the data flow within a VMware VADP solution.
Table 4-11 VMware VADP data flow
Step
number

Data flows from

Data flows to

VADP integrated backup server requests a VM.

vCenter Server via Ethernet

vCenter Server identifies the correct ESX server and requests a snapshot
of the target VM.

ESX server via Ethernet

A software snapshot is created within the ESX server.

Within the ESX server

HPE StoreServ (snapshot image backup data)

Backup server via FC

Backup server

HPE StoreOnce storage or HPE StoreEver tape


library via FC

The backup server sends a command to destroy the backup snapshot.

vCenter Server via Ethernet

The vCenter Server directs the correct ESX host to destroy the backup
snapshot.

ESX server via Ethernet

Veeam backup and restore to HPE StoreEver tape storage with HPE StoreServ
integration

Typical usage:
SAN backup of VMware and Microsoft Hyper-V to a physical tape library.
HPE products offer integration with many third-party software backup solutions, such as Veeam, a
solution that is popular with customers running the VMware virtual infrastructure.
Figure 4.16 illustrates a virtual environment utilizing a Veeam backup solution with the HPE StoreServ
integration. Veeam works directly with the HPE StoreServ storage to create storage snapshots and mount
them to the Veeam Proxy. It is similar to the direct SAN access, but less impactful on the infrastructure
and the VMs being backed up.

Figure 4-16 Veeam backup to tape connectivity and data flow

Table 4.12 explains the data flow within the Veeam backup to tape solution.

Table 4-12 Veeam backup to tape data flow


Step
number

Data flows from

Data flows to

Veeam backup server

ESX server to create a snapshot via LAN/WAN

The ESX server creates the snapshot and sends


acknowledgement.

Veeam Proxy server via LAN (occurs simultaneously with


step 3)

Veeam backup server

HPE StoreServ storage creates the snapshot.

Veeam backup server

The ESX server deletes the snapshot via LAN.

The StoreServ storage snapshot is mounted.

Veeam Proxy server via FC

The Veeam Proxy server stores the data.

Veeam repository server via FC

Veeam repository server

HPE StoreEver tape storage via FC

Figure 4.17 illustrates a VM restore from the HPE StoreEver tape storage with Veeam.

Figure 4-17 VM restore from tape with Veeam connectivity and data flow

Table 4.13 explains the data flow when a VM is restored from tape with Veeam.

Table 4.13 VM restore from tape with Veeam data flow


Step
number

Data flows from

Data flows to

The system administrator initiates the restore operation from the


Veeam backup server via LAN
Veeam GUI.

The Veeam backup server sends a command to load the correct


HPE StoreEver tape library via FC
tapes.

The HPE StoreEver tape library sends the restore image.

Veeam backup server via FC

The Veeam backup server extracts the VM from the restore


image.

Veeam repository server via LAN

The Veeam backup server restores the VM image.

ESX server and datastore LUNs on the HPE StoreServ


storage via LAN

The Veeam backup server sends a command to delete the


restore image.

Repository server via LAN

Figure 4.18 illustrates a user-directed restore from a tape with Veeam.

Figure 4-18 User-directed restore from tape with Veeam connectivity and data flow

Table 4.14 explains the data flow when a user initiates a restore from tape using Veeam.

Table 4-14 User-directed restore from tape with Veeam


Step
number Data flows from

Data flows to

The system administrator initiates a restore from the Veeam GUI.

Veeam backup server


via LAN

The Veeam backup server sends a command to load the correct tapes.

HPE StoreEver tape


library via FC

The HPE StoreEver tape library sends the restore image.

Veeam backup server


via FC

The Veeam backup server sends the restore image.

Veeam repository
server via FC

The system administrator manually directs the VM restore or other restore options such as Microsoft
Exchange or Microsoft SharePoint using the Veeam GUI.

Veeam backup server


via LAN

Veeam backup to HPE StoreOnce with tape offload

Typical usage:
Offload to tape using Veeam backup and replication software.
Figure 4.19 illustrates how HPE StoreEver is used with Veeam backup to tape (archival of data from HPE
StoreOnce). Veeam backup and replication enables the ability to archive backup files stored in backup
repositories to HPE StoreEver tapes.

Figure 4-19 Veeam backup to HPE StoreOnce connectivity and process flow

Within this configuration, the process flow is as follows:


1.The Veeam backup server enumerates backup files in the database backup catalog and tape
catalog to detect if any data was modified since the last backup and queues the changes for
archiving.
2.The Veeam backup server connects to the transport service to start the data transport process.
3.The modified data flows from the backup repository on the HPE StoreOnce VSA to the HPE
StoreEver tape. For this process to occur, the source transport service on the Veeam backup server
reads the data from the HPE StoreOnce VSA and the target transport service on the Veeam backup
server writes the data to the HPE StoreEver tape.
4.The Veeam backup server updates the backup and tape catalogs.

Hpe Enterprise Secure Key Manager (Eskm)


Encryption ensures information security and consistency. Data is encrypted at the source, in transit, and at
the destination. Modern encryption algorithms are based on digital certificates also known as security
keys. Loosing or exposing private keys means losing the information confidentiality; therefore, key
management becomes very important.
As the number of services within the company increases, the number of security keys that must be
managed multiplies. The keys must be kept secure and backed up properly. HPE offers the Enterprise
Secure Key Manager (ESKM) appliance to handle backup and restoration of security keys.

Figure 4-20 HPE Enterprise Secure Key Manager

ESKM is a complete key management solution for generating, storing, serving, controlling, and auditing
access to data encryption keys in a secure appliance. Figure 4.21 shows the integration of encryption tape
appliances into a SAN environment consisting of multiple servers, primary disk storage (HPE StoreServ),
and an HPE StoreEver tape library. The backup data goes directly to the HPE StoreEver LTO tape drive.
The ESKM sends an encryption key to the tape library across the Ethernet and the library passes the key
to the drive via its management port.

Figure 4-21 HPE ESKM connectivity and data flow

Table 4.15 explains the data flow within the HPE ESKM solution.

Table 4-15 HPE ESKM data flow


Step
number

Data flows from

Data flows to

SAN-attached HPE StoreServ storage

SAN-attached server via FC

SAN-attached server

SAN-attached HPE StoreEver tape library via FC

HPE StoreEver tape library requests and receives the key.

ESKM via Ethernet

Library passes the key to the LTO tape drive over the
management port.

ESKM via Ethernet

ESKM (1)

ESKM (2) keys are automatically replicated between the nodes


via Ethernet.

Tiered Data Retention For Hpe Storeserv


Figure 4.22 illustrates a SAN that combines HPE StoreServ storage, HPE StoreOnce backup, and HPE
StoreEver tape to offer a complete business protection solution. Users can implement a disk-based
solution with the HPE StoreOnce backup for daily, short-term backups and fast recovery. HPE StoreEver
tape is used in conjunction with the HPE StoreOnce backup to address long-term archival and compliance
objectives. As the data on the HPE StoreOnce backup appliance becomes infrequently or never accessed,
the tiered data retention policies can be used to expire the data on the HPE StoreOnce while moving that
data to a secondary copy on the HPE StoreEver tape. You can also transport a secondary LTO cartridge
offsite for DR purposes. This reduces risk, increases efficiency, and lowers storage costs, which enable
you to maximize the value you get from your data over its entire lifecycle while minimizing TCO.

Figure 4-22 Tiered data retention for HPE StoreServ connectivity and data flow

Table 4.16 explains how the data flows within the tiered data retention for HPE StoreServ solution.

Table 4-16 Tiered data retention for HPE StoreServ data flow
Step
number

Data flows from

Data flows to

Primary storage (HPE StoreServ


7440c)

SAN-attached server via FC

SAN-attached server

SAN-attached HPE StoreOnce 5100 and SAN-attached HPE StoreEver MSL6480


via FC

SAN-attached HPE StoreEver


MSL6480

Offsite tape vaulting

Learning check questions


Reinforce your knowledge and understanding of the topics just covered by completing this learning check:
1. Which HPE solution consists of a disk-based storage appliance for backing up network media
servers or PCs to target devices on the appliance and supports source and target deduplication?
a)StoreServ Remote Copy
b)StoreOnce Catalyst remote site replication
c)StoreAll backup to StoreEver tape storage with Symantec NetBackup
d)StoreServ File Persona Share Backup
2. Your customer is looking for a backup solution for dispersed remote and branch offices. These
offices run virtualized environments and must send their deduplicated data to a central data center
and its DR site over a WAN connection. Which HPE solution should you recommend to this
customer for these offices?
a)StoreServ File Persona software NDMP backup
b)StoreAll backup to StoreEver tape storage with Symantec NetBackup
c)StoreOnce Catalyst remote site replication
d)StoreOnce VSA
3. What does VADP accomplish?
a)It enables centralized backup software to back up VMs without running individual backups
inside these VMs.
b)It enables automatic data archiving from StoreServ arrays to StoreEver without using
specialized archival software.
c)It generates, stores, serves, controls, and audits access to data encryption keys in a secure
appliance.
d)It transports data between the NAS devices and the backup devices.

Learning check answers


This section contains answers to the learning check questions.

1. Which HPE solution consists of a disk-based storage appliance for backing up network media
servers or PCs to target devices on the appliance and supports source and target deduplication?
a)StoreServ Remote Copy
b)StoreOnce Catalyst remote site replication
c)StoreAll backup to StoreEver tape storage with Symantec NetBackup
d)StoreServ File Persona Share Backup
2. Your customer is looking for a backup solution for dispersed remote and branch offices. These
offices run virtualized environments and must send their deduplicated data to a central data center
and its DR site over a WAN connection. Which HPE solution should you recommend to this
customer for these offices?
a)StoreServ File Persona software NDMP backup
b)StoreAll backup to StoreEver tape storage with Symantec NetBackup
c)StoreOnce Catalyst remote site replication
d)StoreOnce VSA
3. What does VADP accomplish?
a)It enables centralized backup software to back up VMs without running individual
backups inside these VMs.
b)It enables automatic data archiving from StoreServ arrays to StoreEver without using
specialized archival software.
c)It generates, stores, serves, controls, and audits access to data encryption keys in a secure
appliance.
d)It transports data between the NAS devices and the backup devices.

Learner Activity
Using your favorite search engine, search the Internet for the HPE Data Agile BURA Compatibility
Matrix. Save it on your desktop and open it.
Using the compatibility matrix, answer these questions:
1. What is the published data of this document and version?
_____________________________________________________________________________
2. What are the key updates in this version?
_____________________________________________________________________________
_____________________________________________________________________________
_____________________________________________________________________________
_____________________________________________________________________________
3. To back up data to an HPE StoreAll target running OS version 6.5.x, what minimum version of HPE

Data Protector must you use?


_____________________________________________________________________________
4. You want to connect the StoreEver Ultrium 920 (LTO-3 HH SCSI) tape drives to your existing
ProLiant DL servers. Which ProLiant DL servers are not supported with these tape drives?
_____________________________________________________________________________
_____________________________________________________________________________
_____________________________________________________________________________
_____________________________________________________________________________
5. Are VMware ESX hosts supported with direct-attached SAS tape drives? (Hint: Use the Virtual
Machine Support section.)
Yes
No
6. Are VMware ESX hosts supported with direct-attached SCSI tape drives? (Hint: Use the Virtual
Machine Support section.)
Yes
No

Learner activity answers


This section contains answers to the learner activity.
1. What is the published data of this document and version?
April 2015, version 2 (However, this answer will change over time as new documents are released.)
2. What are the key updates in this version? (The answer will also vary, depending on the release of this
document.)
New StoreOnce Software version 3.12.1
New Support for Symantec Backup Exec 15
New LTO-6 Firmware
New MSL6480 Tape Library Firmware 4.70
3. To back up data to an HPE StoreAll target running OS version 6.5.x, what minimum version of HPE
Data Protector must you use (Figure 4.23)?

Figure 4-23 Applications certified to back up data using StoreAll as a target

4. You want to connect the StoreEver Ultrium 920 (LTO-3 HH SCSI) tape drives to your existing
ProLiant DL servers. Which ProLiant DL servers are not supported with these tape drives (Figure
4.24)?

Figure 4-24 HPE StoreEver standalone tape support

5. Are VMware ESX hosts supported with direct-attached SAS tape drives? (Hint: Use the Virtual
Machine Support section.)
Yes
No
6. Are VMware ESX hosts supported with direct-attached SCSI tape drives? (Hint: Use the Virtual
Machine Support section.)
Yes
No

Additional References

An IDC press release: Companies Looking to Cut Cost &$3;mp;amp; Complexity out of Storage
Infrastructure Drive Growth for New Storage Technologies in the Second Quarter
class="ImageCaption">http://www.idc.com/getdoc.jsp?containerId=prUS25883215

Summary
BURA solutions can be direct-attached or SAN-based. HPE Data Agile BURA Compatibility Matrix
provides information for designing data protection solutions.
Simple backup configurations use the point-to-point topology, where the backup device is connected
directly to the backup server. Data is then copied directly to the tape media. HPE StoreEver and
StoreOpen solutions can be used in this topology.
SAN-based backup configurations are more advanced and often employ sophisticated methodologies,
such as connecting independent SAN fabrics, partitioning tape libraries, and extending SAN over long
distances.
Remote backup configurations vary in complexity and in technologies they employ. They include multisite
replication, deduplication, tiered backups and archival, backups of clusters and virtual environments, and
snapshots.
Encryption ensures data security and consistency. Modern encryption algorithms use security keys, which
must be protected and managed. Enterprise Secure Key Manager is a complete key management solution
for generating, storing, serving, controlling, and auditing access to data encryption keys in a secure
appliance.

5 Practice Test
Introduction
The Designing HPE Backup Solutions exam (HPE0-J77) tests your ability to understand technologies,
concepts, and strategies of effective Enterprise Backup Solutions, including identifying the correct
components for a backup and recovery solution. It also tests your understanding of HPE's BURA solutions
to provide customers with complete data protection. This exam also tests your competency level and
understanding of HPE StoreOnce Backup Systems and BURA best practices, configuration guidelines, and
recommended configurations.
The intent of this study guide is to set expectations about the context of the exam and to help candidates
prepare for it. Recommended training to prepare for this exam can be found at the HPE Certification and
Learning website (http://certification-learning.hpe.com) and in books like this one.
It is important to note that although training is recommended for exam preparation, successful completion
of the training alone does not guarantee you will pass the exam. In addition to the training, exam items are
based on knowledge gained from on-the-job experience and other supplemental reference material that
may be specified in this guide.

Minimum Qualifications
Candidates should have a minimum of one year of experience in storage technologies with a strong
understanding of how to design enterprise storage solutions based on customer needs.

HPE0-J77 EXAM DETAILS


The following are details about the HPE0-J77 exam:

Number of items: 50
Exam time: 1 hour 15 minutes
Passing score: 70%
Item types: Multiple choice (single-response), multiple choice (multiple response), matching, and
drag-and-drop
Reference material: No online or hard copy reference material is allowed at the testing site.

HPE0-J77 TESTING OBJECTIVES


This exam validates that you can successfully perform the following objectives. Each main objective is
given a weighting, indicating the emphasis this item has in the exam.
24% Differentiate and apply foundational storage architectures and technologies.
Describe backup systems technology.

Describe data availability methods and technologies.


Describe and differentiate between replication, backup and archiving.
40% Plan and design HPE storage solutions.

Discover backup solution opportunities.


Plan and design the backup solution.
Size the backup solution.
Review and validate the backup solution proposal.
Present the backup solution to the customer.

17% Performance-tune, optimize, and upgrade HPE storage backup solutions.


Identify and compare the existing backup solution design to the best practices documentation.
Optimize and performance-tune the storage backup solution.6% Tune, optimize, and upgrade HP
Networking solutions.
19% Manage, monitor, administer, and operate HPE storage backup solutions.

Install, configure, set up, and demo proof-of-concept for HPE storage backup solutions.
Demonstrate data replication tasks for proof-of-concept.
Demonstrate external array storage for proof-of-concept.
Demonstrate backup and archival storage in a single-site environment for proof-of-concept.

Test Preparation Questions and Answers


The following questions will help you measure your understanding of the material presented in this
eBook. Read all of the choices carefully; there might be more than one correct answer. Choose all correct
answers for each question.

Questions
1. What is the biggest cause of unplanned downtime?
a.Human errors
b.Software bugs and crashes
c.Cyber-attacks and computer viruses
d.Hardware failures
e.Theft
2. Which of the following are examples of intangible costs of downtime? (Select two.)
a.Lost transaction revenue
b.Brand damage
c.Marketing costs

d.Lost inventory
e.Decrease in stock value
3. In your environment, you have 2500 identical disk drives with a MTBF rate of 500,000 per drive.
How frequently can you expect a disk drive failure?
a.Approximately every 2 days
b.Approximately every 4 days
c.Approximately every 8 days
d.Approximately every 16 days
4. What defines the length of time needed to restore the failed system operations back to normal after a
disaster or a disruption of service?
a.RTO
b.MTTR
c.RPO
d.MTBF
5. Which availability technology is ideal for protection against regional and metropolitan disasters?
a.Generators
b.RAID
c.Redundant server components
d.Remote copy
6. During creation of your backup and recovery strategy, which step is recommended by HPE to be
performed after grouping your data into jobs?
a.Calculate the backup frequency
b.Determine the data protection technology
c.Decide on the archival methodology
d.Assign levels of importance to your data
e.Choose your backup tape rotation scheme
7. Your customer has the following backup characteristics and requirements:
o The data set changes frequently each day.
o The customer has only a very limited backup window each day.
o If data is lost and must be restored, the customer can afford a relatively slow restore.
Which type of backup should you recommend?
a.Incremental
b.Differential
c.Image-based
d.Offline
e.File-based
8. What is the biggest disadvantage of the FIFO backup tape rotation scheme?
a.It is too complex.
b.It has the shortest tail of daily backups.
c.It can suffer from possible data loss.

d.It requires the largest number of tapes.


9. Which type of backup, performed by a central backup server, takes the last full backup of multiple
networked clients, appends incremental or differential backups to the full backup set, and then
promotes it to the full backup to save time and network bandwidth?
a.Synthetic backup
b.Virtual full backup
c.Working set backup
d.Block-level incremental backup

10. A customer needs to back up her environment but cannot afford to shut down running business-critical
applications. She is also cost-sensitive to the storage required to protect her environment. Which
technique would best meet this customers requirements?
a.Breakaway mirror snapshots
b.Copy-on-write snapshots
c.Block-level incremental backup
d.Synthetic backup

11. Which technology detects redundant data and stores or transfers only one instance of such data?
a.Deduplication
b.Snapshots
c.Copy-on-write
d.Data tiering
e.Split-mirrors

12. Which term is used to describe a situation in which the deduplication algorithm computes the same hash
number for a new block of data as the data already stored and does not store this new data?
a.True negative
b.False negative
c.True positive
d.False positive

13. Which technology transmits a single ray of light as its carrier and is used for long distances?
a.FCoE
b.iSCSI
c.Multimode fiber
d.Shortwave fiber
e.Single mode fiber

14. Refer to Figure 5.1.

Figure 5-1 Sample tNAS solution

Which component is referred to by label A in this tiered storage and data lifecycle management
solution based on tNAS?
a.QNM agent
b.QStar ASM server
c.QStar ASM cache
d.StoreAll Gateway

15. What is required to implement tier 3 (electronic vaulting) disaster recovery plan?
a.A second, fully operational site (permanently running but not processing)
b.A second, fully operational site (permanently running and actively processing)
c.An automated, business-integrated solution with high-speed replication and clustering
d.Two active sites with high-bandwidth connection between them and two-phase transaction
commit

16. What must you use to connect multiple independent SAN fabrics that consist of HPE StoreServ storage,
HPE StoreEver tape libraries, and HPE StoreOnce VTLs? (Select two.)
a.Multiprotocol router
b.StoreAll Gateway server
c.StoreServ File Persona server
d.Fibre Channel Data-Plane Forwarder
e.FCR-licensed Fibre Channel switch

17. Refer to Figure 5.2.

Figure 5-2 Tape library partitioning

Figure 5-2 shows HPE StoreEver tape library partitioning for concurrent access from applications A
and B. Which software from HPE must be used to achieve this configuration?
a.StoreOpen Enterprise
b.StoreServ Remote Copy
c.Command View for Tape Libraries
d.StoreAll backup to StoreEver tape storage with Symantec NetBackup

18. Refer to Figure 5.3.

Figure 5-3 HPE StoreServ File Persona Share Backup configuration

Which highlighted component within the HPE StoreServ File Persona Share Backup configuration is
responsible for using the appropriate protocol to read the data from the HPE StoreServ File Persona
and writing it to the target backup device?
a.Media server
b.vCenter server
c.ISV NDMP agent
d.Gateway server

19. What is the target use case for HPE StoreOnce Catalyst remote site replication?
a.Data protection of StoreAll Gateways
b.Transfer of data from a branch office to the main data center
c.Automated management of local backup and offsite backup copies
d.Disaster tolerance, business continuity, and long-distance replication

20. What does the NDMP protocol achieve in backup and disaster recovery environments? (Select two.)
a.It removes the need to transport data through the backup server.
b.It enables long-distance replication through incompatible gateways.
c.It securely transports data encryption keys over an unprotected LAN.
d.It enables automatic data archiving from StoreServ arrays to StoreEver without using specialized
archival software.
e.It controls data backup and recovery communications between the primary and secondary storage
in a heterogeneous environment.

Answers
1. What is the biggest cause of unplanned downtime?
a.Human errors

b.Software bugs and crashes


c.Cyber-attacks and computer viruses
d.Hardware failures
e.Theft
Figure 1-1, Causes of downtime and data loss or unavailability, indicates that hardware failures
attribute 39% of unplanned downtime, followed by human errors (29%), software corruption (14%),
theft (9%), and computer viruses (6%).
For more information, see Chapter 1, section Describing causes of downtime.
2. Which of the following are examples of intangible costs of downtime? (Select two.)
a.Lost transaction revenue
b.Brand damage
c.Marketing costs
d.Lost inventory
e.Decrease in stock value
Examples of intangible and indirect costs related to downtime include the following:

Loss of business opportunities


Loss of employees and/or employee morale
Loss of goodwill in the community
Decrease in stock value
Loss of customers and/or departure of business partners
Brand damage
Shift of market share to competitors
o Bad publicity and press

For more information, see Chapter 1, section Estimating the cost and impact of downtime.
3. In your environment, you have 2500 identical disk drives with a MTBF rate of 500,000 per drive.
How frequently can you expect a disk drive failure?
a.Approximately every 2 days
b.Approximately every 4 days
c.Approximately every 8 days
d.Approximately every 16 days
The formula for the total MTBF is as follows:

Therefore,

For more information, see Chapter 1, section Calculating MTBF.


4. What defines the length of time needed to restore the failed system operations back to normal after a
disaster or a disruption of service?

a.RTO
b.MTTR
c.RPO
d.MTBF
RTO defines the length of time necessary to restore the system operations after a disaster or a
disruption of service. RTO includes at minimum the time to correct the situation (break fix) and to
restore any data. It can, however, include factors such as detection, troubleshooting, testing, and
communication to the users.
For more information, see Chapter 1, section Recovery Time Objective.
5. Which availability technology is ideal for protection against regional and metropolitan disasters?
a.Generators
b.RAID
c.Redundant server components
d.Remote copy
Table 1-2, Options that may be used to achieve availability, indicates that remote copy is a
technology that can protect against regional and metropolitan downtimes. The other options can aid in
such protection but are implemented within the servers or data centers and are, therefore, impacted by
regional and metropolitan disasters.
For more information, see Chapter 1, section Availability technologies.
6. During creation of your backup and recovery strategy, which step is recommended by HPE to be
performed after grouping your data into jobs?
a.Calculate the backup frequency
b.Determine the data protection technology
c.Decide on the archival methodology
d.Assign levels of importance to your data
e.Choose your backup tape rotation scheme
Per HPE, your backup and recovery plan should include these ordered steps:
a.Identifying which data must be saved.
b.Grouping this data into jobs.
c.Assigning levels of importance to this data.
d.Calculating the backup frequency.
e.Protecting this data.
f.Testing your backup operations.
g.Maintaining your backup history.
h.Archiving your tapes.
i.Defining and testing your restore procedures
For more information, see Chapter 2, section Backup basics your backup and recovery plan.
7. Your customer has the following backup characteristics and requirements:
o The data set changes frequently each day.
o The customer has only a very limited backup window each day.
o If data is lost and must be restored, the customer can afford a relatively slow restore.

Which type of backup should you recommend?


a.Incremental
b.Differential
c.Image-based
d.Offline
e.File-based
According to recommendations, you should choose an incremental backup when:
o You have a high percentage of daily changes in your data set.
o You only have a limited backup window per day.
o You can accept a relatively slow restore (which translates to a long downtime).
For more information, see Chapter 2, section When to choose incremental vs. differential backup.
8. What is the biggest disadvantage of the FIFO backup tape rotation scheme?
a.It is too complex.
b.It has the shortest tail of daily backups.
c.It can suffer from possible data loss.
d.It requires the largest number of tapes.
The FIFO scheme suffers from the possibility of data loss. Suppose an error is introduced into the data
but the problem is not identified until several generations of backups and revisions have taken place.
When the error is detected, all backup files contain this error.
For more information, see Chapter 2, section First In, First Out (FIFO).
9. Which type of backup, performed by a central backup server, takes the last full backup of multiple
networked clients, appends incremental or differential backups to the full backup set and then
promotes it to the full backup to save time and network bandwidth?
a.Synthetic backup
b.Virtual full backup
c.Working set backup
d.Block-level incremental backup
The question describes a synthetic backup, which takes the last full backup from clients and appends
their incremental or differential backups to the full backup set. Once the appending procedure is done,
the result is stored to the same backup target as the source data sets and is promoted to the full backup.
For more information, see Chapter 2, section Synthetic backup.

10. A customer needs to back up her environment but cannot afford to shut down running business-critical
applications. She is also cost-sensitive to the storage required to protect her environment. Which
technique would best meet this customer s requirements?
a.Breakaway mirror snapshots
b.Copy-on-write snapshots
c.Block-level incremental backup
d.Synthetic backup
The customers need to perform backups without stopping the running applications leads to snapshots
(neither the synthetic backup nor the block-level incremental backup meet this requirement).

Then, the choice is between the breakaway mirror (clone) snapshot and the copy-on-write (point-intime) snapshot. Breakaway mirror has a larger storage overhead because it requires at least two sets
of data. Therefore, copy-on-write is the best solution for this customer.
For more information, see Chapter 2, section Snapshots (frozen images, off-host backups, or
nondisruptive backups).

11. Which technology detects redundant data and stores or transfers only one instance of such data?
a.Deduplication
b.Snapshots
c.Copy-on-write
d.Data tiering
e.Split-mirrors
Per Wikipedia, data deduplication is a specialized data compression technique for eliminating
duplicate copies of repeating data. This is also described in the Deduplication section of Chapter 3.
For more information, see Chapter 3, section Deduplication.

12. Which term is used to describe a situation in which the deduplication algorithm computes the same hash
number for a new block of data as the data already stored and does not store this new data?
a.True negative
b.False negative
c.True positive
d.False positive
When a hash collision occurs, the system will not store the new data because it sees that its hash
number already exists in the index. This is called a false positive and can result in data loss.
For more information, see Chapter 3, section Deduplication.

13. Which technology transmits a single ray of light as its carrier and is used for long distances?
a.FCoE
b.iSCSI
c.Multimode fiber
d.Shortwave fiber
e.Single mode fiber
Per SNIA dictionary (http://www.snia.org/education/dictionary/s), single mode fiber is optical fiber
designed for the transmission of a single ray or mode of light as a carrier. Single mode fiber
transmission is typically used for long-distance signal transmission.
For more information, see the Learner activity section in Chapter 3.

14. Refer to Figure 5.1.

Figure 5-1 Sample tNAS solution

Which component is referred to by label A in this tiered storage and data lifecycle management
solution based on tNAS?
a.QNM agent
b.QStar ASM server
c.QStar ASM cache
d.StoreAll Gateway
The referenced component in Figure 5.1 is QStar ASM cache, which is the tNAS cache behind the
mount point of the tNAS solution. It is used as a temporary storage until the data is written to the tape
drive behind the cache.
For more information, see Chapter 3, section Tiered storage and data lifecycle management with
tNAS.

15. What is required to implement tier 3 (electronic vaulting) disaster recovery plan?
a.A second, fully operational site (permanently running but not processing)
b.A second, fully operational site (permanently running and actively processing)
c.An automated, business-integrated solution with high-speed replication and clustering
d.Two active sites with high-bandwidth connection between them and two-phase transaction
commit
Companies having a tier 3 disaster recovery plan do have a second and fully operational site which is
permanently up and running (but not processing). In tier 3, electronic vaulting is supported for a subset
of critical data. This means transmitting business-critical data electronically to the secondary site
where backups are automatically created. The permanently running hot site increases the cost even
more, but the time for the recovery of business-critical data is significantly reduced.
A second, fully operational site (permanently running and actively processing) option is describing
tier 4 (electronic vaulting with active 2nd site).

Automated, business-integrated solution with high-speed replication and clustering option is


describing tier 7 (highly automated, business-integrated solution).
Two active sites with high-bandwidth connection between them and two-phase transaction commit
is describing tier 5 (two-site, two-phase commit).
For more information, see Chapter 3, section Backup and disaster recovery tiers.

16. What must you use to connect multiple independent SAN fabrics that consist of HPE StoreServ storage,
HPE StoreEver tape libraries, and HPE StoreOnce VTLs? (Select two.)
a.Multiprotocol router
b.StoreAll Gateway server
c.StoreServ File Persona server
d.Fibre Channel Data-Plane Forwarder
e.FCR-licensed Fibre Channel switch
Figure 4-5 in Chapter 4 illustrates a configuration consisting of independent fabrics, with HPE
StoreServ storage, HPE StoreEver tape libraries, and HPE StoreOnce VTLs. To enable
interconnection of devices between these SAN fabrics, without merging them, a multiprotocol router
(MPR) or a Fibre Channel routing-enabled switch is used.
For more information, see Chapter 4, section Connected independent fabrics.

17. Refer to Figure 5.2.

Figure 5-2 Tape library partitioning

Figure 5-2 shows HPE StoreEver tape library partitioning for concurrent access from applications A

and B. Which software from HPE must be used to achieve this configuration?
a.StoreOpen Enterprise
b.StoreServ Remote Copy
c.Command View for Tape Libraries
d.StoreAll backup to StoreEver tape storage with Symantec NetBackup
Utilizing the partitioning feature in the tape library management console and HPE Command View for
Tape Libraries, each backup application or department is presented with a logical library comprised
of a subset of drives and slots from the physical library.
For more information, see Chapter 4, section HPE StoreEver tape library partitioning.

18. Refer to Figure 5.3.

Figure 5-3 HPE StoreServ File Persona Share Backup configuration

Which highlighted component within the HPE StoreServ File Persona Share Backup configuration is
responsible for using the appropriate protocol to read the data from the HPE StoreServ File Persona
and writing it to the target backup device?
a.Media server
b.vCenter server
c.ISV NDMP agent
d.Gateway server
The media server uses the appropriate protocol to read the data from the File Persona and writes it to the
target.
For more information, see Chapter 4, section HPE StoreServ File Persona Share Backup.

19. What is the target use case for HPE StoreOnce Catalyst remote site replication?
a.Data protection of StoreAll Gateways
b.Transfer of data from a branch office to the main data center
c.Automated management of local backup and offsite backup copies

d.Disaster tolerance, business continuity, and long-distance replication


The typical usage of HPE StoreOnce Catalyst remote site replication is transferring data from a branch
office to the main data center or copying data between sites.
For more information, see Chapter 4, section HPE StoreOnce Catalyst remote site replication.

20. What does the NDMP protocol achieve in backup and disaster recovery environments? (Select two.)
a.It removes the need to transport data through the backup server.
b.It enables long-distance replication through incompatible gateways.
c.It securely transports data encryption keys over an unprotected LAN.
d.It enables automatic data archiving from StoreServ arrays to StoreEver without using specialized
archival software.
e.It controls data backup and recovery communications between the primary and secondary
storage in a heterogeneous environment.
NDMP is an open protocol used to control data backup and recovery communications between the
primary and secondary storage in a heterogeneous network environment.
The NDMP protocol transports the data between the NAS devices and the backup devices. This
removes the need for transporting the data through the backup server itself, thus enhancing speed and
removing load from the backup server.
For more information, see Chapter 4, sections HPE StoreAll backup to HPE StoreEver tape storage
with Symantec NetBackup and HPE StoreServ File Persona software NDMP backup.