Quality%20Models%20and%20Quality%20Attributes

1
Quality Models and Quality Attributes
2
Outline

• Process and product quality
• Improving the quality of the process can improve
  the quality of the product.
• Process quality does not address all qualities.
• Specifying quality requirements
• Some quality attributes can be quantifiably
  measured, others have qualitative values which
  are observed.
• Understanding quality models
• A quality model is a taxonomy of quality
  attributes and their relationships.
• Architecting with quality attributes
• Several quality attributes are influenced by the
  architecture.

3
Process and Product Quality

• Process qualities are a measure of an activity
  the end result of which is a product.
• Few quality models for the software architecting
  process exist.
• Product quality is a measure of a tangible
  product itself which could be the final
  executable system or intermediate products like
  specifications, models, plans, etc, which
  collectively form the architectural description.

4
Specifying Quality Requirements

• Sample questions to discover information about
  quality attributes
• How does the application fit into the
  enterprises objectives, processes, and other
  information systems?
• How does it relate to other systems including
  manual business processes?
• Is it intuitive and usable or does it sit in a
  virtual corner collecting cyber-dust?

5
Specifying Quality Requirements (Contd)

• A quality attribute is a property of a process or
  product that can have some qualitative or
  quantitative value and can be measured or
  observed.
• Examples of quality attributes are
• Maintainability
• Portability
• Functionality
• usability

6
Specifying Quality Requirements (Contd)

• A quality requirement is a specification of the
  acceptable values of a quality attribute that
  must be present in the system.
• By leaving certain attributes out of the
  requirements specification, it implies that these
  attributes are not important.
• However, they were probably omitted due to lack
  of understanding of the quality or how to specify
  it.

7
Specifying Quality Requirements (Contd)

• The external view of an applications qualities
  is called its characteristics (usability,
  performance, etc.).
• Internal characteristics or subcharacteristics
  are quality attributes viewable by the developers
  (throughput, load balancing, fault tolerance,
  etc.).
• One internal attribute may influence one or more
  characteristics.
• One characteristic may be influenced by one or
  more internal attributes.

8
Measuring Quality Attributes

• A metric is a qualitative or quantitative
  measurement or scale or a specified quality
  attribute together with a method or technique for
  observing or measuring the quality.
• Examples of internal metrics are the complexity
  of the code logic, or the performance of specific
  functions.
• The most common external metrics are tests of the
  systems functions, reliability, and performance.

9
Measuring Quality Attributes (Contd)

• Quality metrics should be specified as
  unambiguously as possible.
• Qualities may be specified with scenarios
  specified by the software designers and
  architects with assistance form the acquirers and
  users.

10
Quality Requirements and Architectural Design

• Quality measurements are performed primarily on
  the architectural description.
• While some measurements are quantitative, most
  are qualitative.
• Qualitative evaluations, while not as rigorous as
  quantitative evaluations, can be objective and
  provide useful predictions about he success of
  the system in meeting quality requirements.

11
Quality Requirements and Architectural Design
(Contd)

• Quality attributes can be addressed through
  architectural styles and patterns.
• Different architectural styles address different
  sets of quality attributes and to varying
  degrees.
• The specification of quality attributes,
  therefore affects the architectural style of the
  system.

12
Quality Requirements and Architectural Design
(Contd)

• Not all quality attributes are addressed by the
  architectural design.
• Usability and performance both have
  nonarchitectural aspects.

13
Systems Knowledge and Quality Attributes

• In terms of levels of systems knowledge, the
  identification of quality attributes is
  equivalent to the source level (level 0) of
  systems knowledge.
• Quality requirements are specific values or
  ranges of values for these attributes which is
  data level knowledge about the system (level 1).

14
Systems Knowledge and Quality Attributes (Contd)

• Software architecting begins with creating models
  (generative level of systems knowledge, or level
  2) that can generate the data level knowledge.
• We evaluate the architecture to see if it indeed
  produces the same data as specified in level 1.

15
Barriers to Achieving Quality

• Common reasons for failure to meet high levels of
  quality include
• Misunderstanding of the importance of quality
  attributes
• Inadequate languages for expressing and
  specifying quality requirements
• Inadequate modeling methods and notations for
  expressing solutions that address specific
  quality attributes

16
Barriers to Achieving Quality (Contd)

• Common reasons for failure to meet high levels of
  quality include (contd)
• Difficulty in designing for quality attributes
• Lack of documented design and architecture
  patterns for addressing quality attributes
• Quality control as an afterthought in most
  projects

17
Common Quality Attribute Misunderstandings

• Specifying Quality Requirements
• There has been little work in the field of formal
  specification for nonfunctional requirements
• Modeling Methods Dont Address Quality Attributes
• Design methodologies are weak in representing
  nonfunctional requirements
• Designing for Quality Attributes
• Quality attributes are interdependent and cannot
  be achieved in isolation

18
Common Quality Attribute Misunderstandings
(Contd)

• Solution Catalogues and Quality Attributes
• Design patterns can be used to address some
  quality attributes, but mapping them to a design
  patterns context and problem statements is not
  easy.
• It may be worthwhile to begin cataloguing
  patterns with respect to standardized quality
  attributes
• Quality Control is an Afterthought
• Putting testing off until the end of the project
  is effectively a statement that quality is not
  important enough to be assessed early

19
Understanding Quality Models

• A quality model is a specification of the
  required characteristics that a software system
  must exhibit.
• Several quality models have been developed over
  the years.
• McCall/GE Quality Model
• Organizes quality around three uses of software
  product revision, product operations, and product
  transition

20
Understanding Quality Models (Contd)

• Several quality models have been developed over
  the years. (Contd)
• Boehm Quality Model
• Shares a common subset with the McCall model and
  identifies additional quality attributes
• ISO/IEC 9126-12001 Quality Standard
• Identifies six quality characteristics
  functionality, reliability, usability,
  efficiency, maintainability, and portability.
  These are further divided into several
  subcharacteristics

21
Understanding Quality Models (Contd)

• Recent work at the SEI at CMU has produced a
  framework for specifying quality models by adding
  more precision to quality attribute definitions
  and introducing metrics and analysis techniques
  for measuring software and architecture
  descriptions.
• Most development organizations do not employ
  formal metrics for nonfunctional requirements.
  They rely on testing, inspections and subjective
  assessments of quality.
• If done right, these methods can be an effective
  way of predicting system quality.

22
Quality Metamodel
External Characteristic
Observable Via Execution
1
1
manifestation of
Internal Quality Attribute name value
Not Observable Via Execution
Engineering Practice
addresses 1 1
1 measure
1
Architecting Practice
Metric scale method
Scenario
23
The McCall/GE Model Factors

• Factors (characteristics) describe the external
  view of the software as viewed by users of the
  application.
• Product operation factors
• Accuracy
• Reliability
• Efficiency
• Integrity
• Usability

24
The McCall/GE Model Factors (Contd)

• Product revision factors
• Maintainability
• Flexibility
• Testability
• Product transition factors
• Interface facility
• Reusability
• Transferability

25
Additional Factors from Boehm

• Validity
• Clarity
• Understandability
• Modifiability
• Modularity
• Generality
• Economy
• Resilience
• Documentation

26
DeGrace and Stahl Software Engineering Practices

• Coding practices
• Management practices
• Use of design heuristics

27
Laurence Bests Application Architecture

• One of the first attempts at applying quality
  characteristics to architecture
• Previously models did not focus on architecture
  but rather the entire process and product of
  software engineering
• Bests characteristics
• Accuracy and comprehensiveness
• Simplicity of use

28
Laurence Bests Application Architecture (Contd)

• Bests characteristics (contd)
• Operational flexibility
• Ease of development, maintenance, and extension
• Security and control
• Resource efficiency
• Recovery

29
Bass, Clements, and Kazman Categories for Quality
Attributes

• Observable via execution
• Not observable via exceution
• Each of these is divided into three
  subcategories
• System attributes
• Business qualities
• Architecture description qualities

30
Bass, Clements, and Kazman Categories for Quality
Attributes (Contd)

• Observable via execution
• Not observable via execution
• The following attributes may be observed by
  executing the system
• Performance
• Security
• Availability
• Functionality
• Usability

31
Bass, Clements, and Kazman Categories for Quality
Attributes (Contd)

• Usability is composed of six quality attributes
• Learnability
• Efficiency
• Memorability
• Error avoidance
• Error handling
• Satisfaction

32
Bass, Clements, and Kazman Categories for Quality
Attributes (Contd)

• The following are not observed by executing the
  system
• Modifiability
• Portability
• Reusability
• Integrability
• Testability

33
Bass, Clements, and Kazman Categories for Quality
Attributes (Contd)

• Some quality attributes are business related
• Time to market
• Cost
• Targeted market
• Rollout schedule
• Extensive use of legacy systems

34
Bass, Clements, and Kazman Categories for Quality
Attributes (Contd)

• In the SEI quality model, quality attributes have
  three characterizations
• External stimuli,
• Architectural decisions
• Responses
• The requirements need to be specified in such a
  way as to be observable.
• For example, maintainability should be specified
  as concrete change scenarios that a software
  developer would do to modify the system.

35
The ISO/IEC 9126 Standard

• The six characteristics are
• Functionality
• Reliability
• Usability
• Efficiency
• Maintainability
• Portability

36
Benefits of Quality Models

• The ISO/IEC 9126 identifies the following uses of
  a quality model
• Validate the completeness of a requirements
  definition
• Identify software requirements
• Identify software design objectives
• Identify software testing objectives
• Identify user acceptance criteria for a completed
  software product

37
Benefits of Quality Models (Contd)

• A quality model that defines quality attributes
  precisely improves communications between
  acquirers, architects, and developers.
• A quality model that specifies usable and
  practical metrics can improve the quality of
  design models.
• Effective application of a quality model adds
  more quality control earlier in the projects
  life cycle.

38
Architecting with Quality Attributes

• The quality attributes that can be addressed by
  an applications architecture are
• Functionality
• Performance (efficiency)
• Modifiability
• Availability and reliability
• Usability
• Portability

39
Functionality

• It is the ability of the system or application to
  satisfy the purpose for which it was designed.
• It is related to validity, correctness,
  interoperability, and security.
• It drives the initial decomposition of the
  system.
• It is the basis upon which all other quality
  attributes are specified

40
Interoperability

• It is the quality of a system that enables it to
  work with other systems.
• It includes the quality to work with other
  systems not yet known.

41
Security

• It is the ability to enforce authorization,
  authentication, and deliberate denial of service
  attacks.
• Security requirements can affect the functional
  decomposition of the system.-

42
Performance (Efficiency)

• It represents the responsiveness of the system
  (e.g., the time required to respond to events or
  the number of events processed in some period of
  time).
• Some aspects of performance are architectural
  (the distribution of computations across
  components) and some are not (choices of
  algorithms).
• Performance has been a driving factor in systems
  architecture and often compromises the
  achievement of other quality attributes.

43
Resource Efficiency

• One aspect of performance is the efficient
  utilization of resources.
• The decomposition of a system into layers may
  impede the systems ability to satisfy
  performance requirements.

44
Modifiability

• It is the ability to grow an architecture over
  time.
• A modifiable architecture can have new features
  added without requiring architectural rework

45
Availability and Reliability

• Availability is an attribute that measures the
  proportion of time the system is up and running.
• Reliability is an attribute that measures
  the systems ability to
  continue operating over time.
• Both are partially addressed by the architecture.

46
Recoverability

• It is the ability of a system to pick up where it
  left off after a shutdown or crash.
• This might involve launching diagnostic programs
  to check the integrity of the system before the
  actual system launches.

47
Usability

• This typically refers to the usability with
  respect to the end user.
• It can also address other system users such as
  system maintainers, operators, etc.
• It is measured using scenarios.

48
Portability

• It is the ability to reuse a component in a
  different application or operating environment.
• It can be considered as a special kind of
  modifiability.
• Portability related attributes are
• Adaptability
• Installability
• Conformance
• Replaceability

49
Portability (Contd)

• Extensibility is another word for portability.
• One way to satisfy the extensibility requirement
  is to produce a set of platform-independent
  architectural models. This forces the designer
  to consider the architecture from a
  platform-independent point of view.

50
Architecting and Quality Models

• The architectural practices used to address
  quality attributes include the use of patterns
  and heuristics.
• Architectural styles are high-level architecture
  patterns that address a specific set of concerns
  or quality attributes.
• Some views (and models) address particular
  quality attributes and not others. (E.g., use
  cases address functionality but not performance.)

51
Architecting and Quality Models (Contd)

• Some quality attributes cannot be entirely
  evaluated by inspecting the models alone some
  require additional context information such as
  scenarios. (E.g., modifiability can not be
  address without specifying what is to be
  modified.) A set of modification scenarios must
  be supplied.

52
Summary

• Quality attributes are measurable or observable
  characteristics of products and processes.
• The quality requirements specification gives
  quantitative or qualitative values for these
  attributes that must be satisfied by the
  executable system.
• It is cost effective to evaluate models and other
  intermediate products with respect to quality
  attributes.

53
Summary (Contd)

• Not all quality attributes can be satisfied by
  architectural decisions.
• Reasons why the use of quality attributes is not
  common.
• Misunderstanding of their importance
• Inadequate languages for expressing them
• Inadequate specification of quality requirements
  in projects
• Inadequate modeling methods and notations
• Inherent difficulty in designing for quality
  attributes
• Lack of documented design and architectural
  patterns
• The fact that quality control is an afterthought
  on most projects