Software Architecture

Software Architecture
New wine in old bottles?
(i.e., software architecture ≅ global design?,
architect ≅ designer)

SE, Software Engg., Prabhat Gangwar, ©2008 2
Overview
 What is it, why bother?
 Architecture Design
 Viewpoints and view models
 Architectural styles
 Architecture asssessment
 Role of the software architect

The Role of the Architect
client,
users
architect developers
appearance,
behaviour
construction,
co-operation
architectural
design
visualises prescribes
requirements solutions
createsassess assess

Pre-architecture life cycle
requirements
agreement
quality
development
stakeholders
(few)

Characteristics
 Iteration mainly on functional requirements
 Few stakeholders involved
 No balancing of functional and quality
requirements

Adding architecture, the easy way
architecture
detailed design
implementation
requirements
agreement
quality
development
stakeholders
(few)

Architecture in the life cycle
requirements
architecture
quality
agreement
stakeholders
(many)
development

Characteristics
 Iteration on both functional and quality
requirements
 Many stakeholders involved
 Balancing of functional and quality requirements

Why Is Architecture Important?
 Architecture is the vehicle for stakeholder
communication
 Architecture manifests the earliest set of design
decisions
 Constraints on implementation
 Dictates organizational structure
 Inhibits or enable quality attributes
 Architecture is a transferable abstraction of a
system
 Product lines share a common architecture
 Allows for template-based development
 Basis for training

Where did it start?
 1992: Perry & Wolf
 1987: J.A. Zachman; 1989: M. Shaw
 1978/79: David parnas, program families
 1972 (1969): Edsger Dijkstra, program families
 1969: I.P. Sharp @ NATO Software Engineering
conference:
“I think we have something in addition to software
engineering [..] This is the subject of software
architecture. Architecture is different from
engineering.”

Software architecture, definition (1)
The architecture of a software system defines that
system in terms of computational components
and interactions among those components.
(from Shaw and Garlan, Software Architecture,
Perspectives on an Emerging Discipline, Prentice-
Hall, 1996.)

statement
⇓
procedure
⇓
module
⇓
(design) pattern
⇓
architecture

Software Architecture, definition (2)
The software architecture of a system is the
structure or structures of the system, which
comprise software elements, the externally visible
properties of those elements, and the
relationships among them.
(from Bass, Clements, and Kazman, Software
Architecture in Practice, SEI Series in Software
Engineering. Addison-Wesley, 2003.)

 Important issues raised in this definition:
 multiple system structures;
 externally visible (observable) properties of components.
 The definition does not include:
 the process;
 rules and guidelines;
 architectural styles.

Architectural Structures
 module structure
 conceptual, or logical structure
 process, or coordination structure
 physical structure
 uses structure
 calls structure
 data flow
 control flow
 class structure

Software Architecture, definition (3)
Architecture is the fundamental organization of a
system embodied in its components, their
relationships to each other and to the
environment and the principles guiding its design
and evolution
(from IEEE Standard on the Recommended
Practice for Architectural Descriptions, 2000.)

 Architecture is conceptual.
 Architecture is about fundamental things.
 Architecture exists in some context.

Other points of view
 Architecture is high-level design
 Architecture is overall structure of the system
 Architecture is the structure, including the
principles and guidelines governing their design
and evolution over time
 Architecture is components and connectors

Software Architecture & Quality
 The notion of quality is central in software
architecting: a software architecture is devised to
gain insight in the qualities of a system at the
earliest possible stage.
 Some qualities are observable via execution:
performance, security, availability, functionality,
usability
 And some are not observable via execution:
modifiability, portability, reusability, integrability,
testability

Overview

Attribute-Driven Design (Bass et al, Ch 7)
 Choose module to decompose
 Refine this module:
 choose architectural drivers (quality is driving force)
 choose pattern that satisfies drivers
 apply pattern
 Repeat steps

Example ADD iterations
 Top-level: usability ⇒ separate user interface ⇒
Seeheim/three tier architecture
 Lower-level, within user interface: security ⇒
authenticate users
 Lower-level, within data layer: availability ⇒
active redundancy

Generalized model
 Understand problem
 Solve it
 Evaluate solution

Global workflow in architecture design
context
requirements
evaluation
results
architecture
backlog
synthesis
evaluation

Generalized model (cont’d)
backlog
sign.reqs
constraints
ideas
assets
architecture
eval results
synthesis
evaluation

Design issues, options and decisions
A designer is faced with a series of design issues
 These are sub-problems of the overall design problem.
 Each issue normally has several alternative solutions (or
design options)
 The designer makes a design decision to resolve each issue.
 This process involves choosing the best option from
among the alternatives.

Taking decisions
Design
problem
sub-
proble
m
(or
issue)
sub-
proble
m
(or
issue)
Design
option
Desig
n
option
Desig
n
option
Desig
n
option
Problem
space
Decision
space
Alternative
solutions
Alternative
solutions
Decision =
best option
Decision =
best option

Decision space
The space of possible designs that can be achieved
by choosing different sets of alternatives.
client-server
monolithic
fat-client
thin-client
client in a
separate
user interface
layer
no separate
user interface
layer
Programmed in Java
Programmed in Visual Basic
Programmed in C++
client
style

Tree or graph?
 Issues and options are not independent ...
client-server
monolithic
fat-client
thin-client
client in a
separate
user interface
layer
no separate
user interface
layer
client
style
flexibility
layered MVC

More than just IT
 Technical and non-techical issues and options are
intertwined
 Architects deciding on the type of database
versus
 Management deciding on new strategic partnership
or
 Management deciding on budget

Some (tacit) decisions are related to norms
and values

Types of decisions
 Implicit, undocumented
 Unaware, tacit, of course knowledge
 Explicit, undocumented
 Vaporizes over time
 Explicit, explicitly undocumented
 Tactical, personal reasons
 Explicit, documented
 Preferred, exceptional situation

Why is documenting design decisions
important?
 Prevents repeating (expensive) past steps
 Explains why this is a good architecture
 Emphasizes qualities and criticality for
requirements/goals
 Provides context and background

Uses of design decisions
 Identify key decisions for a stakeholder
 Make the key decisions quickly available. E.g., introducing
new people and make them up2date.
 ..., Get a rationale, Validate decisions against reqs
 Evaluate impact
 If we want to change an element, what are the elements
impacted (decisions, design, issues)?
 ..., Cleanup the architecture, identify important architectural
drivers

Elements of a design decision
Issues: design issues being addressed
Decision
Status: e.g., pending, approved
Assumptions: underlying assumptions
Alternatives
Rationale; the why of the decision taken
Implications: e.g. need for further decisions

Pointers on design decisions
 Hofmeister et al, Generalizing a Model of Software
Architecture Design from Five Industrial Approaches,
Journal of Systems and Software, 2007
 Tyree and Ackerman, Architecture decisions: demystifying
architecture, IEEE Software, vol. 22(2), 2005.
 Kruchten, Lago and van Vliet, Building up and exploiting
architectural knowledge, WICSA, 2005.
 Lago and van Vliet, Explicit assumptions enrich
architectural models, ICSE, 2005.

Overview

Software design in UML
 Class diagrams, state diagrams, sequence
diagram, etc
 Who can read those diagrams?
 Which type of questions do they answer?
 Do they provide enough information?

Who can read those diagrams?
 Designer, programmer, tester, maintainer, etc.
 Client?
 User?

Which type of questions do they answer?
 How much will it cost?
 How secure will the system be?
 Will it perform?
 How about maintenance cost?
 What if requirement A is replaced by requirement
B?

Analogy with building architecture
 Overall picture of building (client)
 Front view (client, “beauty” committee)
 Separate picture for water supply (plumber)
 Separate picture for electrical wiring (electrician)
 etc

Architecture presentations in practice
 By and large two flavors:
 Powerpoint slides – for managers, users, consultants, etc
 UML diagrams, for technicians
 A small sample …

So, …
 Different representations
 For different people
 For different purposes
 These representations are both descriptive and
prescriptive

IEEE model for architectural descriptions
Mission
Sy stemEnv ironment Architecture
Rationale
Architecture
Description
Concern
Library
Viewpoint
Viewpoint
Stakeholder
Model
View
Mission
Sy stemEnv ironment Architecture
Rationale
Architecture
Description
Concern
Library
Viewpoint
Viewpoint
Stakeholder
Model
View

Some terms (from IEEE standard)
 System stakeholder: an individual, team, or
organization (or classes hereof) with interests in,
or concerns relative to, a system.
 View: a representation of a whole system from the
perspective of a related set of concerns.
 Viewpoint: A viewpoint establishes the purposes
and audience for a view and the techniques or
methods employed in constructing a view.

Stakeholders
 Architect
 Requirements engineer
 Designer (also of other systems)
 Implementor
 Tester, integrator
 Maintainer
 Manager
 Quality assurance people

Viewpoint specification
 Viewpoint name
 Stakeholders addressed
 Concerns addressed
 Language, modeling techniques

Kruchten’s 4+1 view model
Logical
Viewpoint
Implementation
Viewpoint
Process
Viewpoint
Deployment
Viewpoint
Scenarios
End-user
Functionality
Programmers
Software management
Integrators
Performance
Scalability
System engineers
Topology
Communications

4 + 1: Logical Viewpoint
 The logical viewpoint supports the functional
requirements, i.e., the services the system should
provide to its end users.
 Typically, it shows the key abstractions (e.g.,
classes and interactions amongst them).

4 + 1: Process Viewpoint
 Addresses concurrent aspects at runtime (tasks,
threads, processes and their interactions)
 It takes into account some nonfunctional
requirements, such as performance, system
availability, concurrency and distribution, system
integrity, and fault-tolerance.

4 + 1: Deployment Viewpoint
 The deployment viewpoint defines how the
various elements identified in the logical, process,
and implementation viewpoints-networks,
processes, tasks, and objects-must be mapped
onto the various nodes.
 It takes into account the system's nonfunctional
requirements such as system availability,
reliability (fault-tolerance), performance
(throughput), and scalability.

4 + 1: Implementation Viewpoint
 The imlementation viewpoint focuses on the
organization of the actual software modules in the
software-development environment.
 The software is packaged in small chunks-
program libraries or subsystems-that can be
developed by one or more developers.

4 + 1: Scenario Viewpoint
 The scenario viewpoint consists of a small subset
of important scenarios (e.g., use cases) to show
that the elements of the four viewpoints work
together seamlessly.
 This viewpoint is redundant with the other ones
(hence the "+1"), but it plays two critical roles:
 it acts as a driver to help designers discover architectural elements
during the architecture design;
 it validates and illustrates the architecture design, both on paper and
as the starting point for the tests of an architectural prototype.

Architectural views from Bass et al
(view = representation of a structure)
 Module views
 Module is unit of implementation
 Decomposition, uses, layered, class
 Component and connector (C & C) views
 These are runtime elements
 Process (communication), concurrency, shared data (repository),
client-server
 Allocation views
 Relationship between software elements and environment
 Work assignment, deployment, implementation

Module views
 Decomposition: units are related by “is a
submodule of”, larger modules are composed of
smaller ones
 Uses: relation is “uses” (calls, passes information
to, etc). Important for modifiability
 Layered is special case of uses, layer n can only
use modules from layers <n
 Class: generalization, relation “inherits from”

Component and connector views
 Process: units are processes, connected by
communication or synchronization
 Concurrency: to determine opportunities for
parallelism (connector = logical thread)
 Shared data: shows how data is produced and
consumed
 Client-server: cooperating clients and servers

Allocation views
 Deployment: how software is assigned to
hardware elements
 Implementation: how software is mapped onto file
structures
 Work assignment: who is doing what

How to decide on which viewpoints
 What are the stakeholders and their concerns?
 Which viewpoints address these concerns?
 Prioritize and possibly combine viewpoints

Decision visualization

Business
viewpoint

A caveat on quality
 A view can be used to assess one or more quality
attributes
 E.g., some type of module view can be used to
assess modifiability
 It should then expose the design decisions that
affect this quality attribute

Overview

Architectural styles
 An architectural style is a description of
component and connector types and a pattern of
their runtime control and/or data transfer.
 Examples:
 main program with subroutines
 data abstraction
 implicit invocation
 pipes and filters
 repository (blackboard)
 layers of abstraction

Alexander’s patterns
 There is abundance evidence to show that high buildings
make people crazy.
 High buildings have no genuine advantage, except in
speculative gains. They are not cheaper, they do not help to
create open space, they make life difficult for children, they
wreck the open spaces near them. But quite apart from this,
empirical evidence shows that they can actually damage
people’s minds and feelings.
 In any urban area, keep the majority of buildings four stories
high or less. It is possible that certain buildings should
exceed this limit, but they should never be buildings for
human habitation.

General flavor of a pattern
 IF you find yourself in <context>, for example
<examples>, with <problem>
 THEN for some <reasons>, apply <pattern> to
construct a solution leading to a <new context>
and <other patterns>

Components and Connectors
 Components are connected by connectors.
 They are the building blocks with which an
architecture can be described.
 No standard notation has emerged yet.

Types of components
 computational: does a computation of some sort.
E.g. function, filter.
 memory: maintains a collection of persistent data.
E.g. data base, file system, symbol table.
 manager: contains state + operations. State is
retained between invocations of operations. E.g.
adt, server.
 controller: governs time sequence of events. E.g.
control module, scheduler.

Types of connectors
 procedure call (including RPC)
 data flow (e.g. pipes)
 implicit invocation
 message passing
 shared data (e.g. blackboard or shared data base)
 instantiation

Framework for describing architectural
styles
 problem: type of problem that the style addresses.
Characteristics of the reqs’s guide the designer in
his choice for a particular style.
 context: characteristics of the environment that
constrain the designer, req’s imposed by the
style.
 solution: in terms of components and connectors
(choice not independent), and control structure
(order of execution of components)
 variants
 examples

Main-program-with-subroutines style
problem: hierarchy of functions; result of functional
decomposition, single thread of control
context: language with nested procedures
solution:
 system model: modules in a hierarchy, may be weak or strong,
coupling/cohesion arguments
 components: modules with local data, as well as global data
 connectors: procedure call
 control structure: single thread, centralized control: main program
pulls the strings
variants: OO versus non-OO

Abstract-data-type style
problem: identify and protect related bodies of information.
Data representations likely to change.
context: OO-methods which guide the design, OO-languages
which provide the class-concept
solution:
 system model: component has its own local data (= secret it hides)
 components: managers (servers, objects, adt’s)
 connectors: procedure call (message)
 control structure: single thread, usually; control is decentralized
variants: caused by language facilities

Implicit-invocation style
 problem: loosely coupled collection of components. Useful
for applications which must be reconfigurable.
 context: requires event handler, through OS or language.
 solution:
 system model: independent, reactive processes, invoked when an
event is raised
 components: processes that signal events and react to events
 connectors: automatic invocation
 control structure: decentralized control. Components do not know
who is going to react.
 variants: Tool-integration frameworks, and languages with
special features.

Pipes-and-filters style
 problem: independent, sequential transformations on
ordered data. Usually incremental, Ascii pipes.
 context: series of incremental transformations. OS-
functions transfer data between processes. Error-handling
difficult.
 solution:
 system model: continuous data flow; components incrementally
transform data
 components: filters for local processing
 connectors: data streams (usually plain ASCII)
 control structure: data flow between components; component has
own flow
 variants: From pure filters with little internal state to batch
processes

Repository style
problem: manage richly structured information, to be
manipulated in many different ways. Data is long-lived.
context: shared data to be acted upon by multiple clients
solution:
system model: centralized body of information. Independent
computational elements.
components: one memory, many computational
connectors: direct access or procedure call
control structure: varies, may depend on input or state of computation
variants: traditional data base systems, compilers, blackboard
systems
Shared
Data
Client
Client
Client

Layered style
problem: distinct, hierarchical classes of services.
“Concentric circles” of functionality
context: a large system that requires decomposition (e.g.,
virtual machines, OSI model)
solution:
 system model: hierarchy of layers, often limited visibility
 components: collections of procedures (module)
 connectors: (limited) procedure calls
 control structure: single or multiple threads
variants: relaxed layering
Layern
Layer2
Layer1

Model-View-Controller (MVC) style
problem: separation of UI from application is desirable due to
expected UI adaptations
context: interactive applications with a flexible UI
solution:
 system model: UI (View and Controller Component(s)) is decoupled
from the application (Model component)
 components: collections of procedures (module)
 connectors: procedure calls
 control structure: single thread
variants: Document-View
Model
ViewController
nn

Overview

Architecture evaluation/analysis
 Assess whether architecture meets certain quality
goals, such as those w.r.t. maintainability,
modifiability, reliability, performance
 Mind: the architecture is assessed, while we hope
the results will hold for a system yet to be built

Software Architecture Analysis
Software
architecture
System
Properties Qualities
implementation
properties

Analysis techniques
 Questioning techniques: how does the system
react to various situations; often make use of
scenarios
 Measuring techniques: rely on quantitative
measures; architecture metrics, simulation, etc

Scenarios in Architecture Analysis
 Different types of scenarios, e.g. use-cases, likely
changes, stress situations, risks, far-into-the-
future scenarios
 Which stakeholders to ask for scenarios?
 When do you have enough scenarios?

Preconditions for successful assessment
 Clear goals and requirements for the architecture
 Controlled scope
 Cost-effectiveness
 Key personnel availability
 Competent evaluation team
 Managed expectations

Architecture Tradeoff Analysis Method
(ATAM)
 Reveals how well architecture satisfies quality
goals, how well quality attributes interact, i.e. how
they trade off
 Elicits business goals for system and its
architecture
 Uses those goals and stakeholder participation to
focus attention to key portions of the architecture

Benefits
 Financial gains
 Forced preparation
 Captured rationale
 Early detection of problems
 Validation of requirements
 Improved architecture

Participants in ATAM
 Evaluation team
 Decision makers
 Architecture stakeholders

Phases in ATAM
0: partnership, preparation (informally)
1: evaluation (evaluation team + decision makers,
one day)
2: evaluation (evaluation team + decision makers +
stakeholders, two days)
3: follow up (evaluation team + client)

Steps in ATAM (phase 1 and 2)
 Present method
 Present business drivers (by project manager of system)
 Present architecture (by lead architect)
 Identify architectural approaches/styles
 Generate quality attribute utility tree (+ priority, and how
difficult)
 Analyze architectural approaches
 Brainstorm and prioritize scenarios
 Analyze architectural approaches
 Present results

Example Utility tree
Utility
Performance
Usability
Maintainability
Transaction response time (H, M)
Throughput
Training
150 transactions/sec
Database vendor releases
new version
Normal operations

Outputs of ATAM
 Concise presentation of the architecture
 Articulation of business goals
 Quality requirements expressed as set of
scenarios
 Mapping of architectural decisions to quality
requirements
 Set of sensitivity points and tradeoff points
 Set of risks, nonrisks, risk themes

Important concepts in ATAM
 Sensitivity point: decision/property critical for
certain quality attribute
 Tradeoff point: decision/property that affects
more than one quality attribute
 Risk: decision/property that is a potential problem
 These concepts are overlapping

Software Architecture Analysis Method
(SAAM)
 Develop scenarios for
 kinds of activities the system must support
 kinds of changes anticipated
 Describe architecture(s)
 Classify scenarios
 direct -- use requires no change
 indirect -- use requires change
 Evaluate indirect scenarios: changes and cost
 Reveal scenario interaction
 Overall evaluation

Scenario interaction in SAAM
 Two (indirect) scenarios interact if they require
changes to the same component
 Scenario interaction is important for two reasons:
 Exposes allocation of functionality to the design
 Architecture might not be at right level of detail

Overview

Role of the software architect
 Key technical consultant
 Make decisions
 Coach of development team
 Coordinate design
 Implement key parts
 Advocate software architecture

Summary
 new and immature field
 proliferation of terms: architecture - design
pattern - framework - idiom
 architectural styles and design pattern describe
(how are things done) as well as prescribe (how
should things be done)
 stakeholder communication
 early evaluation of a design
 transferable abstraction

Further reading
 Mary Shaw and David Garlan, Software Architecture;
Perspectives of an Emerging Discipline, 1995.
 Philippe B. Kruchten, The 4+1 view model of architecture,
IEEE Software, 12(6):42-50, November 1995.
 Frank Buschmann et al., Pattern-Oriented Software
Architecture: A System of Patterns, 1996. Part II: 2001.
 Erich Gamma et al., Design Patterns: Elements of Reusable
Object-Oriented Software, 1995.
 Len Bass et al, Sofware Architecture in Practice, 2003 (2nd
edition).
 C. Hofmeister et al., Applied Software Architecture, 1999.
 Jan Bosch, Design & Use of Software Architectures, 2000.

Software Architecture

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Software Architecture (20)

More from Prabhat gangwar (20)

Recently uploaded (20)

Software Architecture

Editor's Notes