Architecture Quality Attributes
From Suhrid.net WikiJump to navigationJump to search
- 1 Intro
- 2 Quality Attribute Scenarios
- 3 Observable Quality Attributes
- 4 Unobservable Quality Attributes
- 5 Business Qualities
- 6 Quality of Architecture
- Qualities that must be accommodated in a systems architecture over and above basic functionality.
- Functionality and quality attributes are mostly independent.
- Functionality can be achieved in any number of ways. It is mostly independent of structure.
- We need to choose the right way/structure so that quality attributes are met.
- Categories of quality attributes:
- Observable attributes. e.g:
- Unobservable attributes. e.g:
- Business attributes.e.g:
- Time to market
- Overall cost
- Backward compatibility
- Observable attributes. e.g:
Quality Attribute Scenarios
- Meaningless to say System is modifiable, reusable etc. How do we justify it ?
- Quality Attributes are best characterised via scenarios.
- SEI format for QA scenarios:
- Source of stimulus: The entity (human, computer system) that has generated the stimulus.
- Stimulus: The stimulus is a condition that needs to be considered when it arrives at a system.
- Environment: The prevailing condition when the stimulus occurs (e.g. system may be overloaded when stimulus occurs or maybe the system is in normal operating conditions)
- Artifact: Some piece of the system is stimulated. This may be the whole system or some pieces of it.
- Response: The response is the activity undertaken after the arrival of the stimulus.
- Response Measure: When the response occurs, it should be measurable so that the requirement can be tested.
- General QA scenarios: These scenarios are system independent and can be applied to any system.
- Specific QA scenarios: These scenarios are derived from the general scenario by instantiating each part. They are specific to the particular system under consideration.
- For e.g. there is a general availability scenario with a range of values that each part (source, stimulus, env etc) can take. From this we can derive concrete, system specific scenarios.
Observable Quality Attributes
- Availability is concerned with system failure and its associated consequences.
- A fault may become a failure if not corrected or masked.
- Availability of a system is the probability that it will be operational when it is needed.
- Defined as (MTTF)/(MTTF + MTTR). Where MTTF = Mean Time to Failure and MTTR = Mean Time to Repair.
- Source: Internal or external to the system.
- Stimulus: The stimulus is a fault. The fault can be omission (fails to respond), crash (repeatedly suffers omission faults), timing (late or early response), response itself (incorrect)
- Artifact: The part of the system that is required to be highly available (processor, communication channel, storage etc)
- Environment: Normal/Degraded.
- Response: One or more of : Logging, Notifying, Disable source of events, become unavailable, continue to operate in normal or degraded mode.
- Response Measure: Time interval when the system must be available or a repair time.
- Is all about timing. Basically concerned with how long it takes the system to respond when an event occurs.
- Source: Internal or external to the system.
- Stimulus: Periodic/Sporadic/Stochastic events arrive.
- Artifact: System
- Environment: Normal mode; Overload mode
- Response: Stimuli must be processed (events handled); change in level of service (normal to degraded)
- Response Measure: Latency - DL by which the event must be processed, Jitter, Throughput, Miss rate, Data loss.
- Security is a measure of the system's ability to resist unauthorized usage while still providing services to legitimate users.
- Security can be characterized by a system providing:
- Non-repudiation : One cannot deny that an action took place.
- Confidentiality: Data is protected from unauthorized access.
- Integrity: Data is delivered as intended, no modification in between.
- Assurance: Parties are who they claim they are.
- Availability: System will be available for legitimate use e.g. DoS will not prevent.
- Auditing: System tracks activities.
- Source: Human or system who can be correctly/incorrectly identified, who is internal/external with access to some resources.
- Stimulus: View data, Change Data, access system services, reduce availability to system services.
- Artifact: System services, System data
- Environment: Online/offline, connected/disconnect, behind firewall, open to network.
- Response: Authenticates user, allows access, blocks access, grants/withdraws permissions, records access, informs a user or another system.
- Response measure: Time/effort/resources required by an attacker to bypass security measures, probability of detecting attack, probability of identifying individual responsible, extent to which data/services are damaged, restoring data/services after attack.
- Concerned with how easy it is for the user to accomplish a desired task and kind of support the system provides.
- Things like: (Basically usability principles)
- Learning system features, Using a system efficiently, minimize impact of errors, adapt system to user needs, increase confidence and satisfaction.
- Source: Always the end user.
- Stimulus: Wants to use system efficiently, learn system features, minimize impact of errors, adapt system, feel comfortable.
- Artifact: Is always the system.
- Environment: At runtime or at system configuration time.
- Response: Provides user with one or more:
- Support learn system: - Context sensitive help, familiar interface
- Support use system efficiently: - Provide shortcuts, search, reuse of commands etc.
- Minimize impact of errors: Support cancel, undo, abort, recover password etc
- Adapt System: I18N, Customizability.
- Feel comfortable: Display system state, work at the user's pace.
- Response Measure: Task time, number of errors, user satisfaction, gain of user knowledge etc.
Unobservable Quality Attributes
- Modifiability is about the cost of change. Two aspects:
- Which artifact can change ? : functions, platforms, capacity, UI etc..
- When is the change made and who makes it ? Source code by Developer, UI by end user, Compile time changes, Run time flags, settings etc, Configuration parameters.
- Source: The person making the changes - Developer, Sys Admin, End User
- Stimulus: Desire to add/edit/delete functionality, quality attribute, capacity.
- Artifact: System UI, functionality, platform, environment
- Environment: At runtime, compile time, build time, design time.
- Response: Locates place where change can be made, makes modification independently, tests modification, deploys modification.
- Response Measure: Cost in terms of no. of elements, quality attributes effected, effort, money etc.
- Ease through which software can be made to demonstrate its faults through testing.
- For a system to be testable, it must be possible to control each components internal state and inputs and then to observe its outputs.
- Source: Unit developer, Integration/System/client acceptance tester, System user.
- Stimulus: Completion of Analysis, architecture, design, class, system increment.
- Artifact: Piece of design, piece of code, complete application.
- Environment: At design time, compile time, deployment time.
- Response: Provide access to state values, provides computed values.
- Response Measure: Percent executable statements executed, time to perform tests, Probability of failure if fault exists.
- How easily can a system be moved to a different target environment.
- Usual strategy: decouple provider from user.
- However interfaces take time to define and add run-time penalty.
- Who pays ? Current project manager might be reluctant, if benefits are only for future projects.
- How easily different parts of the system work together ?
- Common interfaces in terms of data structures and protocols.
- Linked to portability and integrability.
- Reuse within the same system as well as across systems.
- Depends on degree of coupling.
- Time to market: Development time, pressure to buy COTS. Will be easier if system is decomposed into elements.
- Cost and benefit: Architecture that is highly flexible or relies on technology will be more expensive to build.
- Projected lifetime of the system: Long life systems need to have high modifiability, portability, scalability.
- Targeted market: If mass market, the platforms on which a system runs determine size. Thus, portability is key. Product line approach should be considered.
- Rollout schedule: If product introduced in increments, then flexibility, customizability are key.
- Integration with legacy systems: Integrability is important.
Quality of Architecture
- Conceptual Integrity: is the underlying theme or vision that unifies the design of the system at all levels. e.g. Built for massive scalability where scalability is the underlying theme.
- Better to leave out features if they are not central to theme.
- Correctness and completeness:
- Essential to allow all system requirements and resource constraints to be met.
- Buildability: System needs to be completed by available team in a timely manner and be open to changes.
- Tasks easily split into development teams.