CON7247 Oracle Database In-Memory The Next Big Thing

Safe
Harbor Statement
The following is intended to outline our general product direcGon. It is intended for
informaGon purposes only, and may not be incorporated into any contract. It is not a
commitment to deliver any material, code, or funcGonality, and should not be relied upon
in making purchasing decisions. The development, release, and Gming of any features or
funcGonality described for Oracles products remains at the sole discreGon of Oracle.
Copyright 2014, Oracle and/or its aliates. All rights reserved. |
Oracle Database In-Memory

The Next Big Thing
Juan Loaiza
Senior Vice President
Systems Technology
Released July 2014
Oracle Database In-Memory Goals

Real Time
Analytics
Accelerate Mixed
Workload OLTP
No Changes to
Applications
Trivial to
Implement
100x
Row Format Databases vs. Column Format Databases

SALES
Rows Stored
ConGguously
Columns
Stored
ConGguously
TransacCons run faster on row format

Example: Query or Insert a sales order
Fast processing few rows, many columns
SALES
AnalyCcs run faster on column format

Example : Report on sales totals by region
Fast accessing few columns, many rows
OLTP Example
SALES
Row
Query a single sales order in row format

One contiguous row accessed = FAST
Query
SALES

Stores
Query a sales order in Column Format

Many column accessed = S L O W
Column
Query
Query
Query
Query
UnCl Now Must Choose One Format and Suer Tradeos

Breakthrough: Dual Format Database

Normal
Buer Cache
New In-Memory
Format
SALES
SALES
Row
Format
Column
Format
BOTH row and column

formats for same table
Simultaneously acGve and
transacGonally consistent
AnalyGcs & reporGng use new
in-memory Column format
OLTP uses proven row format
SALES
Now we can run Gme-sensiGve

analyGcal queries directly against
our OLTP database. This is something
we wouldn't have dreamt of earlier.
Arup Nanda, Enterprise Architect
Starwood Hotels and Resorts
Oracle In-Memory Columnar Technology

Pure In-Memory Columnar
Pure in-memory column format

Not persistent, and no logging
Quick to change data: fast OLTP

Enabled at table or parGGon

Only acGve data in-memory
SALES

2x to 20x compression typical
Available on all hardware
plaeorms

Early User - Schneider Electric

Global Specialist in Energy Management
25 billion revenue

160,000+ employees in 100+ countries

Tuesday Sept. 30th at 12:00 pm

Francois Bermond from Schneider Electric
is parGcipaGng in session CON 7248
Top Five Things to Know About

10
Schneider In-Memory Compression

Schneider General Ledger Compression Factors
Over 2 billion
General Ledger
Entries
20
Default
15
10
5
8.6x
13x
16x
19x
900 GB on disk
0
Query Low
Query High
Capacity Low
Capacity High
11
Orders of Magnitude Faster AnalyGc Data Scans

Memory
REGION
Example:
Find sales in
region of CAlifornia
CPU
Scans use super fast SIMD

vector instructions
CA
Vector Register
Load
mulGple
region
values
CA
CA
Each CPU core scans local

in-memory columns
Vector
Compare
all values
an 1 cycle
CA
> 100x Faster
Originally designed for

graphics & science
Billions of rows/sec scan

rate per CPU core
Row format is millions/sec
12
Joining and Combining Data Also DramaGcally Faster

Example: Find total sales in outlet stores
Sales
Converts joins of data in

multiple tables into fast
column scans
Type=Outlet
Amount
StoreID in
15, 38, 64
Store ID
Store ID
Type
Stores
Joins tables 10x faster

Sum
13
Generates Reports Instantly

Example: Report sales of footwear in outlet stores
Products
In-Memory
Report Outline
Sales
Dynamically creates in-memory
report outline
Then report outline lled-in
Footwear
Stores
Outlets
Footwear
$$
$$$
during fast fact scan
Reports run much faster

Without predened cubes
Also ooads report ltering to

Outlets
Sales
Exadata Storage servers
14
Schneider Speedup Across 1545 Queries

7x to 128x faster
100
90
80
Seconds per Query

Buer Cache
70
60
50
1545 queries
IN-MEMORY
Currently take 34
hours to complete
CombinaGon of lter
queries, aggregaGons
and summaGons
40
30
20
10
0
2000M 300M 30M
2 billion General
Ledger Entries
5M 0.5M
Million rows returned by query
15
Schneider Speedup vs. Disk

From 62x to 3259x faster
Seconds per Query

900
Disk
800
700
600
500
IN-MEMORY
400
300
200
100
0
2000M 300M
30M 5M 0.5M
Million rows returned
16
Oracle Database In-Memory is a

game changer for OLTP, DW, and
mixed workloads. It dramaGcally
improves the performance of all
types of analyGcal queries.
Liviu Horn, AVP Database Management
McKesson Health SoluGon

17
Complex OLTP is Slowed by AnalyGc Indexes

Table
1 3
OLTP
Indexes
10 20
AnalyGc
Indexes
Most Indexes in complex OLTP

(e.g. ERP) databases are only
used for analytic queries
Inserting one row into a table
requires updating 10-20 analytic
indexes: Slow!
Indexes only speed up predictable
queries & reports

18
Column Store Replaces AnalyGc Indexes

Table
1 3
OLTP
Indexes
In-Memory
Column Store
Fast analytics on any columns

Better for unpredictable analytics
Less tuning & administration
Column Store not persistent so

update cost is much lower
OLTP & batch run faster

19
Schneider Update TransacGons Speedup

4 0 00
3 5 00
Data Sales Accounts
Primary Index Only
3 0 00
2 5 00
Main table has

1 Primary Key +
21 secondary indexes
Primary Index Plus In-Memory Columns
No Index
2 0 00
No Index + IM
1 5 00
Full Index
Test - 303 million

transacGons
1 0 00
500
All Indexes
75
70
65
60
55
50
45
40
35
30
25
20
15
10
Millions of transaction per day
From 5x to 9x faster
Currently takes 21
hours
Millions of records in the target table
20
Schneider Storage ReducGon

Over 70% reducCon in storage usage due to analyCc index removal
Size in GBs
SECONDARY
INDEXES
350
300
250
200
150
TABLES &
PK INDEXES
100
50
0
Objects
Redo
21
Scale-Out In-Memory Database to Any Size

Scale-Out across servers to
grow memory and CPUs
In-Memory queries parallelized
across servers to access local
column data
Direct-to-wire InfiniBand
protocol speeds messaging on
Engineered Systems

22
Full support for RAC scale-out

means Oracle Database In-Memory
can be used on our largest Data
Warehouse, enabling more near
real-Gme analyGcs.
Sudhi Vijayakumar, Senior Oracle DBA
Yahoo Inc.

23
In-Memory Speed + Capacity of Low Cost Disk

DRAM
PCI FLASH
Howest Data
Size not limited by memory

Data transparently accessed
across tiers
AcGve Data
Each tier has specialized

algorithms & compression
Simultaneously Achieve:
DISK
Cold Data
Speed
I/Os
Cost
of DRAM
of Flash
of Disk
24
Scale-Up for Maximum In-Memory Performance

M6-32
Big Memory Machine
32 TB DRAM
32 Socket
3 Terabyte/sec Bandwidth
Scale-Up on large SMPs

Algorithms NUMA optimized
SMP scaling removes

overhead of distributing
queries across servers
Memory interconnect far
faster than any network
25
Oracle In-Memory: Industrial Strength Availability

Data Guard & GoldenGate
Pure In-Memory format does not

change Oracles storage format,
logging, backup, recovery, etc.
RAC
All Oracles proven availability

technologies work transparently
ASM
Protection from all failures

Node, site, corruption, human error,
etc.
RMAN

26
Oracle Database In-Memory: Unique Fault Tolerance

Similar to storage mirroring
Duplicate in-memory
columns on another node
Enabled per table/partition
E.g. only recent data
Application transparent
Downtime eliminated by
using duplicate after failure
Only Available on Engineered Systems

27
DownGme is extremely costly for our

business. Oracles In-Memory
architecture takes the right approach
to balancing real-Gme speed with
conGnuous availability.
Jens-ChrisGan Pokolm
Analyst IT-DB Architecture & Engineering
Postbank Systems AG

28
Oracle In-Memory: Simple to Implement

1. Configure Memory Capacity
inmemory_size = XXX GB
2. Configure tables or partitions to be in memory
alter table | partition inmemory;
3. Later drop analytic indexes to speed up OLTP
29
In terms of how easy the in-memory

opGon was to use, it was actually
almost boring. It just worked
just turn it on, select the tables,
nothing else to do.
Mark Riwman
Chief Technical Ocer
Riwman Mead
30
Oracle In-Memory Requires Zero Application Changes
Full FuncConality
Easy to Implement
Fully CompaCble
- ZERO restricCons on SQL

- No migraGon of data
- All exisGng applicaGons run unchanged
Fully MulCtenant
- Oracle In-Memory is Cloud Ready
Uniquely Achieves All In-Memory Benets With No ApplicaCon Changes
31

made our slowest nancial
queries faster out-of-the box;
then we dropped indexes and
things just got faster.
Evan Goldberg
Co-Founder, Chairman, CTO
NetSuite Inc.
32
We see clear benet from the

Oracle In Memory for our users.
Our exisGng applicaGons were
transparently able to take
advantage of them and no
applicaGon code changes were
required
Scow VanValkenburgh, SAS
33
Oracle ApplicaGons In-Memory Examples

Oracle ApplicaCon Module
Improvement
Elapsed Time
In-Memory Cost Management
1003x Faster
58 hours to 3.5 mins
In-Memory - Financial Analyzer
1,354x Faster
4.3 hours to 11 seconds
In-Memory Sales Order Analysis
1,762x Faster
22.5 minutes to < 1 sec
Subledger Period Close Excep?ons
200x Faster
600 seconds to 3 secs
Call Center Ad-hoc query paBern
1247x Faster
129 seconds to < 1 secs
34
Gezng The Most From In-Memory

Understand Where it Helps

Fast cars speed up travel, not meeGngs

In-Memory speeds up analyGc data access, not:
Network round trips, logon/logo
Parsing, PL/SQL, complex funcGons
Data processing (as opposed to access)
Complex joins or aggregaGons where not much data is ltered before processing
Load and select once Staging tables, ETL, temp tables
Know your boeleneck!

35
Gezng The Most From In-Memory

The Driver Maeers
Avoid stop and go trac

Process data in sets of rows in the Database
Not one row at a Gme in the applicaGon
Plan ahead, take shortest route

Help the opGmizer help you: Gather representaGve set of staGsGcs using
DBMS_STATS
Use all your cylinders

Enable parallel execuGon
In-Memory removes storage bowlenecks allowing more parallelism
36
In-Memory Use Cases
OLTP
Real-Gme reporGng directly on
OLTP source data
Removes need for separate ODS
Speeds data extracGon
Data Warehouse
Staging/ETL/Temp not a candidate
Write once, read once
All or a subset of FoundaGon Layer

For Gme sensiGve analyGcs
PotenGal to replace Access Layer

37
Become a Real-Time Enterprise

Using Oracle Database In-Memory
Data-Driven
DATADRIVEN
AGILE
Get immediate answers to any quesCon

with real-Cme analyCcs
Agile
Eliminate latency with analyCcs directly

on OLTP data
Ecient
EFFICIENT
Easily and Non-disrupCve deployment

accelerates all applicaCons
38
Summary: Oracle Database In-Memory

Powering the Real-Time Enterprise
DATADRIVEN
AGILE
Extreme Performance: AnalyCcs

& OLTP
Extreme Scale-Out & Scale-Up
Extreme Availability
Extreme Simplicity
EFFICIENT
39
Oracle Database In-Memory Schedule for Oracle Open World

Date
Title
LocaCon
Speaker
Monday Sept 29th

13:00 13:45
Oracle Database In-Memory:

The Next Big Thing
Moscone South
Room 103
Juan Loaiza
Senior Vice President, Oracle
Tuesday Sept 30th

12:00 12:45

Moscone South
Room 104
Maria Colgan with

Schneider Electric & Yahoo
Wednesday Oct
1st 14:00 14:45
Oracle Database In-Memory:

Under the Hood
Moscone South
Room 104
Tirthankar Lahiri
Vice President, Oracle
Thursday Oct 2nd

09:30 10:15

Moscone South Maria Colgan with

Room 104
Schneider Electric & Yahoo
Monday 16:15 Oracle Database In-Memory Boot

Tuesday 18:45 Camp: Everything You Need to Get
Wednesday 16:15 Started
Thursday 13:00
Hotel Nikko
Maria Colgan &
Room Peninsula Andy Rivenes

CON7247 Oracle Database In-Memory The Next Big Thing

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

CON7247 Oracle Database In-Memory The Next Big Thing

Hochgeladen von

Copyright:

Verfügbare Formate

Safe

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Oracle Database In-Memory

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Oracle Database In-Memory Goals

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Row Format Databases vs. Column Format Databases

TransacCons run faster on row format

AnalyCcs run faster on column format

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Query a single sales order in row format

Query a sales order in Column Format

UnCl Now Must Choose One Format and Suer Tradeos

Breakthrough: Dual Format Database

BOTH row and column

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Now we can run Gme-sensiGve

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Oracle In-Memory Columnar Technology

Pure in-memory column format

Enabled at table or parGGon

Early User - Schneider Electric

Tuesday Sept. 30th at 12:00 pm

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Schneider In-Memory Compression

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Orders of Magnitude Faster AnalyGc Data Scans

Scans use super fast SIMD

Each CPU core scans local

> 100x Faster

Originally designed for

Billions of rows/sec scan

Joining and Combining Data Also DramaGcally Faster

Converts joins of data in

Joins tables 10x faster

Generates Reports Instantly

Dynamically creates in-memory

Then report outline lled-in

during fast fact scan

Reports run much faster

Also ooads report ltering to

Exadata Storage servers

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Schneider Speedup Across 1545 Queries

Seconds per Query

2000M 300M 30M

Million rows returned by query

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Schneider Speedup vs. Disk

Seconds per Query

Million rows returned

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Oracle Database In-Memory is a

Copyright 2014, Oracle and/or its aliates. All rights reserved. |

Complex OLTP is Slowed by AnalyGc Indexes

Most Indexes in complex OLTP

Column Store Replaces AnalyGc Indexes

Fast analytics on any columns

Column Store not persistent so

Schneider Update TransacGons Speedup

Data Sales Accounts

Primary Index Only

Main table has