architectural design u establishing the overall structure ...bmitchell/course/mcs451/ch_13.pdf ·...
TRANSCRIPT
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 1
Architectural Design
u Establishing the overallstructure of a software system
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 2
Objectivesu To introduce architectural design and its role in
the software processu To describe a number of different types of
architectural modelu To show how the architecture of a system may be
modelled in different waysu To discuss how domain-specific reference models
may be used to compare software architectures
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 3
Topics coveredu System structuringu Control modelsu Modular decompositionu Domain-specific architectures
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 4
Architectural parallelsu Architects are the technical interface between the
customer and the contractor building the systemu A bad architectural design for a building cannot
be rescued by good construction; the same is truefor software
u There are specialist types of building andsoftware architects
u There are schools or styles of building andsoftware architecture
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 5
Architectural design processu System structuring
• The system is decomposed into several principal sub-systemsand communications between these sub-systems are identified
u Control modelling• A model of the control relationships between the different parts
of the system is established
u Modular decomposition• The identified sub-systems are decomposed into modules
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 6
Sub-systems and modulesu A sub-system is a system in its own right whose
operation is independent of the services providedby other sub-systems.
u A module is a system component that providesservices to other components but would notnormally be considered as a separate system
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 7
Architectural modelsu Structure, control and modular decomposition
may be based on a particular model orarchitectural style
u However, most systems are heterogeneous in thatdifferent parts of the system are based ondifferent models and, in some cases, the systemmay follow a composite model
u The architectural model used affects theperformance, robustness, distributability andmaintainability of the system
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 8
System structuringu Concerned with decomposing the system into
interacting sub-systemsu The architectural design is normally expressed as
a block diagram presenting an overview of thesystem structure
u More specific models showing how sub-systemsshare data, are distributed and interface with eachother may also be developed
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 9
Packing robot control systemVisionsystem
Objectidentification
system
Armcontroller
Grippercontroller
Packagingselectionsystem
Packingsystem
Conveyorcontroller
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 10
The repository modelu Sub-systems must exchange data. This may be
done in two ways:• Shared data is held in a central database or repository and may
be accessed by all sub-systems• Each sub-system maintains its own database and passes data
explicitly to other sub-systems
u When large amounts of data are to be shared, therepository model of sharing is most commonlyused
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 11
CASE toolset architecture
Projectreposi tory
Designtranslator
Programeditor
Designeditor
Codegenerator
Designanalyser
Reportgenerator
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 12
Repository model characteristicsu Advantages
• Efficient way to share large amounts of data• Sub-systems need not be concerned with how data is produced
Centralised management e.g. backup, security, etc.• Sharing model is published as the repository schema
u Disadvantages• Sub-systems must agree on a repository data model. Inevitably a
compromise• Data evolution is difficult and expensive• No scope for specific management policies• Difficult to distribute efficiently
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 13
Client-server architectureu Distributed system model which shows how data
and processing is distributed across a range ofcomponents
u Set of stand-alone servers which provide specificservices such as printing, data management, etc.
u Set of clients which call on these servicesu Network which allows clients to access servers
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 14
Film and picture library
Catalogueserver
Catalogue
Videoserver
Fi lm cl ipfiles
Pictureserver
Digitizedphotographs
Hypertextserver
Hypertextweb
Client 1 Client 2 Client 3 Client 4
Wide-bandwidth network
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 15
Client-server characteristicsu Advantages
• Distribution of data is straightforward• Makes effective use of networked systems. May require cheaper
hardware• Easy to add new servers or upgrade existing servers
u Disadvantages• No shared data model so sub-systems use different data
organisation. data interchange may be inefficient• Redundant management in each server• No central register of names and services - it may be hard to
find out what servers and services are available
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 16
Abstract machine modelu Used to model the interfacing of sub-systemsu Organises the system into a set of layers (or
abstract machines) each of which provide a set ofservices
u Supports the incremental development of sub-systems in different layers. When a layerinterface changes, only the adjacent layer isaffected
u However, often difficult to structure systems inthis way
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 17
Version management system
Operatingsystem
Database system
Object management
Version management
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 18
Control modelsu Are concerned with the control flow between
sub-systems. Distinct from the systemdecomposition model
u Centralised control• One sub-system has overall responsibility for control and starts
and stops other sub-systems
u Event-based control• Each sub-system can respond to externally generated events
from other sub-systems or the system’s environment
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 19
Centralised controlu A control sub-system takes responsibility for
managing the execution of other sub-systemsu Call-return model
• Top-down subroutine model where control starts at the top of asubroutine hierarchy and moves downwards. Applicable tosequential systems
u Manager model• Applicable to concurrent systems. One system component
controls the stopping, starting and coordination of other systemprocesses. Can be implemented in sequential systems as a casestatement
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 20
Call-return model
Routine 1.2Routine 1.1 Routine 3.2Routine 3.1
Routine 2 Routine 3Routine 1
Mainprogram
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 21
Real-time system control
Systemcontroller
Userinterface
Faulthandler
Computationprocesses
Actuatorprocesses
Sensorprocesses
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 22
Event-driven systemsu Driven by externally generated events where the
timing of the event is outwith the control of thesub-systems which process the event
u Two principal event-driven models• Broadcast models. An event is broadcast to all sub-systems.
Any sub-system which can handle the event may do so• Interrupt-driven models. Used in real-time systems where
interrupts are detected by an interrupt handler and passed tosome other component for processing
u Other event driven models include spreadsheetsand production systems
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 23
Broadcast modelu Effective in integrating sub-systems on different
computers in a networku Sub-systems register an interest in specific
events. When these occur, control is transferredto the sub-system which can handle the event
u Control policy is not embedded in the event andmessage handler. Sub-systems decide on eventsof interest to them
u However, sub-systems don’t know if or when anevent will be handled
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 24
Selective broadcasting
Sub-system1
Event and message handler
Sub-system2
Sub-system3
Sub-system4
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 25
Interrupt-driven systemsu Used in real-time systems where fast response to
an event is essentialu There are known interrupt types with a handler
defined for each typeu Each type is associated with a memory location
and a hardware switch causes transfer to itshandler
u Allows fast response but complex to program anddifficult to validate
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 26
Interrupt-driven control
Handler1
Handler2
Handler3
Handler4
Process1
Process2
Process3
Process4
Interrupts
Interruptvector
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 27
Modular decompositionu Another structural level where sub-systems are
decomposed into modulesu Two modular decomposition models covered
• An object model where the system is decomposed intointeracting objects
• A data-flow model where the system is decomposed intofunctional modules which transform inputs to outputs. Alsoknown as the pipeline model
u If possible, decisions about concurrency shouldbe delayed until modules are implemented
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 28
Object modelsu Structure the system into a set of loosely coupled
objects with well-defined interfacesu Object-oriented decomposition is concerned with
identifying object classes, their attributes andoperations
u When implemented, objects are created fromthese classes and some control model used tocoordinate object operations
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 29
Invoice processing system
IssueSend reminderAccept paymentSend receipt
invoice #dateamountcustomer
Invoice
customer #nameaddresscredi t period
Customer
invoice #dateamountcustomer #
Receipt
invoice #dateamountcustomer #
Payment
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 30
Data-flow modelsu Functional transformations process their inputs to
produce outputsu May be referred to as a pipe and filter model (as
in UNIX shell)u Variants of this approach are very common.
When transformations are sequential, this is abatch sequential model which is extensively usedin data processing systems
u Not really suitable for interactive systems
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 31
Invoice processing system
Read issuedinvoices
Identi fypayments
Issuereceipts
Findpayments
due
Receipts
Issuepaymentreminder
Reminders
Invoices Payments
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 32
Domain-specific architecturesu Architectural models which are specific to some
application domainu Two types of domain-specific model
• Generic models which are abstractions from a number of realsystems and which encapsulate the principal characteristics ofthese systems
• Reference models which are more abstract, idealised model.Provide a means of information about that class of system andof comparing different architectures
u Generic models are usually bottom-up models;Reference models are top-down models
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 33
Generic modelsu Compiler model is a well-known example
although other models exist in more specialisedapplication domains• Lexical analyser• Symbol table• Syntax analyser• Syntax tree• Semantic analyser• Code generator
u Generic compiler model may be organisedaccording to different architectural models
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 34
Compiler model
Lexicalanalysi s
Syntacticanalysi s
Semanticanalysi s
Codegeneration
Symboltable
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 35
Language processing system
Syntaxanalyser
Lexicalanalyser
Semanticanalyser
Abstractsyntax tree
Grammardefini tion
Symboltable
Outputdefini tion
Pretty-printer
Editor
Optimizer
Codegenerator
Reposi tory
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 36
Reference architecturesu Reference models are derived from a study of the
application domain rather than from existingsystems
u May be used as a basis for systemimplementation or to compare different systems.It acts as a standard against which systems can beevaluated
u OSI model is a layered model for communicationsystems
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 37
OSI reference model
Appl ication
Presentation
Session
Transport
Network
Data link
Physical
7
6
5
4
3
2
1
Communica tions medium
Network
Data link
Physical
Appl ication
Presentation
Session
Transport
Network
Data link
Physical
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 38
Key pointsu The software architect is responsible for deriving
a structural system model, a control model and asub-system decomposition model
u Large systems rarely conform to a singlearchitectural model
u System decomposition models include repositorymodels, client-server models and abstractmachine models
u Control models include centralised control andevent-driven models
©Ian Sommerville 1995 Software Engineering, 5th edition. Chapter 13 Slide 39
Key pointsu Modular decomposition models include data-flow
and object modelsu Domain specific architectural models are
abstractions over an application domain. Theymay be constructed by abstracting from existingsystems or may be idealised reference models