Beruflich Dokumente
Kultur Dokumente
Use a Data Flow Diagram (DFD) to show the relationships among the
business processes within an organization to:
• external systems,
• external organizations,
• customers,
• other business processes.
Method
Data flow diagrams are used to describe how the system transforms
information. They define how information is processed and stored and
identify how the information flows through the processes.
Examples
• End User,
• Purchasing Department,
• Inventory System.
FLOWS
Definition
Flows define the interfaces between the components within the system, and the
system and its external components.
Types of Flows
Flows that transport data around the data flow diagram are called data flows.
Data flows are the pipelines through which data are transmitted between
any two components on a DFD. The composition of data is known and
defined in a data dictionary.
• purchase order,
• customer profile,
• account number,
• product.
Data flows must be given a name that describes the content of the data
being transmitted and a description of the data flow listing the data
elements. For example:
Bi-directional data flows are used only when the information flowing in
each direction is identical.
Multiple data flows can have the same name on a DFD. When this
occurs, the data transmitted in all of the flows with the same name must
be identical.
Data flows are described as complex when there is more than one data
flow going in the same direction between any two entities.
Name the complex data flow to encompass the contents of all the data
flowing along the pipeline.
Trivial Data Flows
Label all data flows with care for they indicate the interface requirements
for a process. Use descriptive names for labelling data flows.
Ensure that all flows to external entities on lower level diagrams balance
with data flows from external entities on upper level diagrams. In other
words, a new data flow cannot be created at a lower level if it was not
identified on an upper level diagram.
Double-headed data flows are permitted when the data flowing in both
directions are identical.
Data flows from/to external entities to/from data stores are not permitted.
All data must flow through a process.
Flows that are used to synchronize the system control flow and do not require
data are called event flows. Common uses include starting, stopping, and
changing the system status. Event flows are also known as control flows.
A data flow name should be a singular noun phrase (e.g., delivery list, customer
information, credit limit).
An event flow name should be a singular verb or noun phrase (e.g., start, stop,
item available, processing complete).
If a data flow contains multiple attributes (i.e., data elements), an attribute list
should be provided.
STORES
Definition
Stores represent information (i.e., data or control) at rest. Stores are used when
different processes need to share information but are active at different times.
Information can be written to a store and read from a store.
Types of Stores
Data stores are created to store information for later use. They represent
data that is temporarily at rest between processes. For example, a data
store is needed to store data that is generated on a daily basis but is
required for a process that runs weekly.
Local data stores are data stores whose accesses are contained
completely within the boundary of a DFD.
A data store can be drawn on the same DFD more than once to reduce
the visual complexity of the drawing. An extra vertical bar is drawn in the
duplicated data stores to indicate that it appears elsewhere on the same
diagram.
Stores that contain status or control information are called event stores. Event
stores are used to save events that have occurred but have not yet been used.
Unlike a data store, an event store has behaviour associated with it which is not
apparent when looking at the data flow diagram. If the system accesses an event
store and the event has not occurred, the system will be suspended until the
event occurs. Once an event has occurred, accessing it will remove it from the
event store.
Guidelines
The data stores on the data flow diagram map to the entities on the entity-
relationship diagram.
Minimize the use of event stores on the DFD. Try using flags instead.
PROCESSES
Definition
Processes are also known as data transforms. Processes transform input flows
into output flows in a defined manner.
A process is a distinct activity (or set of activities) described by its inputs and
outputs. A process describes a unique behaviour that has a beginning and an
end. A process is performed repeatedly.
Types of Processes
A data process transforms input data into output data. Data processes act
directly on data, either from flows or stores. They represent the functionality of
the system.
A control process transforms input events into output events and is used on a
data flow diagram to indicate the presence of a state transition diagram. Control
processes cannot transform data but can control processes that do. A state
transition diagram describes the behaviour of the control process.
Guidelines
A process name and number must be unique. A process can only exist once on a
data flow diagram.
A process must have at least one input and one output. In other words, a process
is asked (i.e., triggered) to do something and then must deliver (i.e., respond).
A data process cannot input or output discrete event flows (i.e., a data process
should not control the system, it should do the work).
A control process cannot input or output data flows (i.e., a control process
should control the system, not do the work).
A process name should start with an active verb (e.g., Produce Items, Control
Production).
A process exists to do something (i.e., transform input flows into output flows),
therefore a process must have an incoming set of requirements to which it must
conform.
An nth-level (i.e., next level of detail) data flow diagram must describe its parent
process. An nth-level data flow diagram is a decomposition of its parent process
and cannot introduce new functionality.
Notation
Extensions
Identify the external events (from the event list) that have a data event
classification. Data events cause data to be transformed. A data flow diagram is
used to describe this type of behaviour.
The event response can be used to determine the required behaviour and name
of a process. Responses and processes are not necessarily at the same level of
abstraction. Responses that represent low-level activities must be grouped into
related processes. Responses that represent high-level functions must be
decomposed into low-level processes.
Group related responses together. Responses can be related by the activities that
they perform (i.e., if responses do the same things you may want to group them
together). Responses can also be related by the data that they use (i.e., if
responses use the same data you may want to group them together).
Different processes should also be created when different external events have
similar responses but process different data.
The same process may be used by different responses as long as the process's
behaviour and data requirements are the same for each event.
You can also use external events to group related processes (i.e., identify parent
processes).
Identify control events (i.e., the events that use a state transition diagram to
describe the system's response). State transition diagrams describe processing
sequence, therefore the processes controlled by a state transition diagram should
be grouped together with the control process that prompts (i.e., controls) them.
Identify the data processes controlled by a control process. Group the data
processes and the control process together under the same parent process.