An Introduction to Software Architecture

Introduction

Informal Definition of SA First step in developing the solution Overall (high level) structure of the software system Software architecture = Components + Connectors What are components and connectors?

A simple example Component 1 Component 2 Component 1.1 Database

Why Software Architecture? Complexity → Divide and Conquer Process: Divide design process to phases Architectural design Detailed design Product: Decompose system to components Assuring fulfillment of required quality attributes (performance, changeability, etc) from the beginning

Roots of Software Architecture Software architecture is similar to building architecture in many ways. The idea is not new. Concepts related to SA have been in the literature since 60’s and 70’s (e.g. modularity, info. hiding). However, the term is new. During the past 10 years SA have received considerable attention and have been subject to many research projects.

Architecture Business Cycle

Motivation We add a new role to software development team: Software Architect What does software architect do? Simply drawing the some diagrams? What else is related to SA? Are 2 SAs developed in different environmental conditions for a single system the same? This part covers two issues: What influences software architecture? What are influenced by software architecture?

Who influences SA?

Customers and End Users Requirements (including qualities such as performance, maintainability, etc) Budget Limitation Time Limitation Force to apply specific technology, methodology, or organizational discipline

Developing Organization Concerns Business issues investing in, and then amortizing the infrastructure (domain analysis rather than application analysis) keeping cost low simplicity of implementation Organizational issues using the current organizational structure utilizing personnel

Technical Environment Current trends: today’s information system are web-based and use middleware systems (e.g. J2EE, .Net) Available technology: decisions on using a centralized or decentralized system depend on processor cost and communication speed; both are changing quantities.

Architect’s Background Architects develop their mindset from their past experiences. Prior good experiences will lead to replication of prior designs. Prior bad experiences will be avoided in the new design.

Summary: Influences on the Architect

Architecture Influences the Development Organization Organizational Structure and Recourses Work units are organized around architectural units Schedule Budget Enterprise Goals Expertise in building a kind of systems Success in a market Evaluating a market Product-line assets

Architecture Influences Customer Requirements Knowledge of customers to ask for particular features in next systems. Support of upgrade, adaptation, etc.

Architecture Influences the Architect’s Experience and Technical Environment Creation of a system affects the architect’s background. Occasionally, a system or an architecture will affect the technical environment.

Architecture Business Cycle

Process Steps in Architecture-Based Development Understanding the requirements Creating, customizing, or selecting the architecture Representing and communicating the architecture Analyzing or evaluating the architecture Implementing based on architecture Ensuring conformance

What is Software Architecture

Is this diagram an architecture? (ATM Software) Control Card Interface Cash Dispenser Keyboard Interface

What are ambiguities in the previous diagram? Nature of the elements (process, class, object, module, function, processor, or etc) Responsibility of elements Type of connections (calls, invokes, uses, signals, sends data, controls, sub-class) Significance of layout Run-time operation of system

Definition The software architecture of a program or computing 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.

Externally Visible vs. Internal Properties of Component Externally visible properties are what assumption other elements can make of an element Provided services (and interface to access those services) Performance Fault handling Shared resource usage … SA intentionally abstracts away internal properties of elements (to better encounter complexity)

Some Points Every software system has an architecture (SA ≠ specification of SA) Specification of architecture can comprise more than one structure Behavior of elements and relationships are defined in SA (abstractly) The definition do not talk about Good and Bad architectures: SA evaluation methods In the literature: Component = Element Connector = Relationship

What issues are architectural? Architectural issues are those issues that are important to us at the SA abstraction level. An issue is architectural if it is not internal to any element. e.g.: Performance is an architectural quality attribute Behavior of an element is architectural to the extent that influences how other elements must be developed

Architectural Pattern An architectural pattern is a description of element and relation types together with a set of constraints on how they may be used. example: client-server, layered, data-centered unresolved issues of a pattern: Exact number of elements and relations during application Behavior of elements during application Configuration (Topology) during application Patterns define constraint on architecture but are not architecture themselves. Patterns are abstraction for a set of architectures.

Reference Model A reference model is a division of functionality together with data flow between pieces. A standard division of a known problem (a mature domain) into parts. e.g. compiler and DBMS Reference model is not architecture

Reference Architecture A reference architecture is a reference model mapped onto software elements and data flows between the components Elements cooperatively implements the functionality defined in the reference model Reference architecture is not the final architecture but is not far from it

Relationship of the previous concepts Reference Model Reference Architecture Software Architecture System Architectural Pattern

Why is architecture important? Handling complexity Communication among stakeholders Requirements and concerns of stakeholders Time Budget Other Resources

Why is architecture important? (cont) Early Design Decisions Constraints implementation and implementers Organizational structure Enables predicting and ensuring quality attributes Makes it possible to reason about and manage change Helps evolutionary prototyping (risk reduction)

Why is architecture important? (cont) SA is a transferable, reusable model Software product lines Component-based development Automatic generation of lower-level models A basis for training A run-time model in self-adaptive and reconfigurable systems

Hazards With regards to SA changes are categorized to Local (a single component) Non-local (a few components) Architectural (architectural style) Once decided architecture is extremely hard to change It impossible to reach to some quality attribute if architecture disallows

Software Arch. vs. System Arch. System Arch. is the overall architecture of system including hardware and software architecture In assuring quality attributes the architect needs to think about system architecture too (e.g. performance or reliability) But architect has more freedom in software architecture than hardware (hardware choices is less under the architects control)

Architectural Structures and Views In construction, there are blueprints of Plan Different sides of construction Electrical wiring Plumbing … Each of these views specifies a single entity (i.e. the construction) from a different perspective (used by a different person, for a different goal). Similarly there are different structures and views in SA.

Structures and Views (cont) Structures is a set of coherent elements and the relations among them. For each structure these we can specify: Types of elements Types of relations A set of syntactic constraints Semantics of the diagram Rationale, principles, and guidelines For what purposes it is useful View is a representation of software architecture based on an structure as written by the architect and read by stakeholders (an instance of the structure) SA is documented by a number of views.

Categorization of Structures Module Structures Component and Connector Structures Allocation Structures Categories are orthogonal

1 Module Structures Elements: modules (units of implementation). Modules are a code based way of considering the system Specifies: Functional responsibility of modules Other elements a module is allowed to use Generalization and specialization relations Run-time operation of software is not a concern from this view

1.1 Decomposition Structure Elements: modules in a hierarchy Relations: is a sub-module of, shares secret with Function Examples: Contributes to system's modifiability, by ensuring that likely changes fall within the scope of at most a few small modules. Often used as the basis for the development project's organization: the structure of the documentation, and its integration and test plans.

1.2 Uses Structure Elements: modules, procedures, or resources on the interfaces of modules Relations: uses: one unit uses another if the correctness of the first requires the presence of a correct version (not a stub of) of the second. Function Example: Allows incremental development

1.3 Layered Structure Is a subclass of uses structure Elements: layers: a coherent set of related functionality Relations: uses (ideally layer n may only use the services of layer n – 1), provides abstraction to Function Example: Layers are often designed as abstractions (virtual machines) that hide implementation specifics below from the layers above, engendering portability.

1.4 Class Structure Elements: classes Relations: inherits from, is an instance of Function Example: Allows us to reason about reuse and the incremental addition of functionality

2 Component and Connector Structures Elements: run-time components (principal units of computation) and connectors (communication vehicle among components.) Specifies: Major executing components and how they interact Major shared data-stores Which part of system is replicated Flow of data through the system What parts can run in parallel How can system structure change as it executes

2.1 Process Structure Elements: processes or threads Relations: attachment (that allow communication, synchronization, and/or exclusion operations) Function Example: Engineering a system's execution performance and availability.

2.2 Shared Data or Repository Structure Elements: data stores, data producers, and data consumers Relations: data-flow Function Example: To ensure good performance and data integrity.

2.3 Client-Server Structure Elements: clients and servers Relations: protocols and message passing infrastructure. Function Example: Separation of concerns (supporting modifiability) Load balancing (supporting runtime performance)

3 Allocation Structures Show the relationship between the software and the elements in one or more external environment in which software is created and executed. Specifies: The processor that executes each software element The file that stores each software element during development Assignments of software to development team

3.1 Deployment Structure Shows how software is assigned to hardware Elements: software (usually a process from a component and connector view), hardware entities, and communication pathways Relations: is-allocated-to and migrates-to (for dynamic allocations) Function Example: Allows reasoning about performance, data integrity, availability, and security.

3.2 Implementation Structure Shows how software elements (usually modules) are mapped to the file structure(s) in the system's development, integration, or configuration management environments. Elements: any logical unit (e.g. module) Relations: implemented in Function Example: management of development activities and build process

3.3 Work Assignment Structure Assigns responsibility for implementing and integrating the modules to the appropriate development teams Elements: any logical unit (e.g. module) Relations: is assigned to Function Example: The architect will know the expertise required on each team The means for factoring functional commonalities and assigning them to a single team, rather than having them implemented by everyone who needs them.

Notes Each structure is useful on its own right but not all structures are used in all projects. Structures are not independent and must be considered together e.g. relationship of modules with components (many to many) Some structures may be the same in some systems Some structures may be combined (e.g. all component and connector structures may be combined in a single structure)

4 + 1 View Model of Architecture Logical: objects and classes (a module view) Process: (a component and connector view) Development: modules, libraries, subsystems, and units of development (an allocation view) Physical: mapping of elements to hardware and communication (an allocation view) Scenarios (Use-cases) view is not itself architectural.

Quality Attributes

Traditional Classification of Requirements Functional Non-Functional (Quality Attributes) A popular software myth: first we build a software that satisfies functional requirements, then we will add or inject non-functional requirements to it. This idea leads to loss of resources and finally poor quality. So we should design for qualities from the very beginning (architecture level).

Functionality and Architecture Functionality and quality attrs are orthogonal [in theory]. But not all qualities are achievable to any level desired with any functionality. Functionality may be achieved in many ways (it is not so architectural.) Architecture is a means of achieving quality attributes by structuring functionality into elements.

Architecture and Qualities Achieving qualities must be considered throughout design (including SA), implementation, and deployment. Qualities have both architectural and non-architectural aspects. For example In usability: selecting form elements vs. supporting undo operation Performance: amount of communication among components vs. algorithms

Architecture and Qualities Quality attributes are not independent and may not be achieved in isolation. Positive Correlation; e.g. Modifiability and Buildability (in many cases) Negative Correlation (conflict); e.g. Reliability vs. Security-The most secure system has the fewest points of failure—typically a security kernel. The most reliable system has the most points of failure—typically a set of redundant processes or processors where the failure of any one will not cause the system to fail. Performance vs. All Other Qualities

Classification of Quality Attributes Qualities of the system: availability, modifiability, performance, security, testability, and usability. Business qualities (such as time to market) that are affected by the architecture. Architecture qualities, such as conceptual integrity.

System Quality Attributes System quality attributes have been of interest to the software community at least since the 1970s Shortcoming of the previous work: The definitions for an attribute are not operational. Modifiability with regards to which aspect? Which quality a particular aspect belongs to. Is a system failure an aspect of availability, an aspect of security, or an aspect of usability? Each attribute community has developed its own vocabulary. Performance community events, security community attacks, availability community failures, and usability community user input may actually refer to the same occurrence.

Classifications of System Quality Attributes Observable via Execution e.g. performance and security Not observable via execution e.g. modifiability and testability The categories are totally independent (orthogonal), although members of the second category indirectly affect members of the first. Non-observable qualities are important too. (sometimes even more important!!!)

Quality Attribute Scenarios Is the solution to the stated problems. A QAS is a quality-attribute-specific requirement, that consists of: Source of stimulus: actuator; e.g. a human or computer system Stimulus: event. Environment: the condition under which the stimulus occurs; e.g. system is overloaded. Artifact: pieces of system that is stimulated. Response: desired reaction Response measure: response should be measurable in some fashion so that the requirement can be tested. Scenarios may be general or concrete (for specific system)

Example: Availability General Scenario

Example: Availability Concrete Scenario

Modifiability Concrete Scenario

Notes on Scenarios Concrete scenarios role for quality attribute requirements is similar to use cases role for functional requirements. A collection of concrete scenarios can be used as the quality attribute requirements for a system. One of the uses of general scenarios is to enable stakeholders to communicate.

System Quality Attributes Availability (related to Reliability) Modifiability (includes Protability and Reusability, Scalability) Performance Security Testability Usability (includes Self-Adaptability and User-Adaptability)

Business Qualities Time to market Cost and benefit Predicted lifetime of the system Targeted market Rollout schedule(ارائه محصول به بازار) Integration with legacy systems

Qualities of the Architecture Conceptual Integrity Conceptual integrity is the most important consideration in system design. It is better to have a system omit certain anomalous features and improvements, but to reflect one set of design ideas, than to have one that contains many good but independent and uncoordinated ideas. [Brooks 75] Correctness and Completeness Buildability There should be a way to evaluate SA

Achieving Qualities

The Whole Story Business Requirements Quality Requirements Tactics Selection Tactics Implementation: Design Patterns & Architectural Patterns

Tactics A tactic is a design decision that influences a quality attribute. e.g. using redundancy to increase availability Tactics can be refined to other tactics to become more concrete; e.g. redundancy: redundancy of data + process

Availability Tactics

Modifiability Tactics

Patterns Ripple effect: اثر بازدارنده Intermediary:واسط،میانجی

Patterns Binding at runtime means that the system has been prepared for that binding and all of the testing and distribution steps have been completed. Deferring binding time also supports allowing the end user or system administrator to make settings or provide input that affects behavior.  

Patterns Runtime registration supports plug-and-play operation at the cost of additional overhead to manage the registration. Publish/subscribe registration, for example, can be implemented at either runtime or load time. Configuration files are intended to set parameters at startup. Polymorphism allows late binding of method calls. Component replacement allows load time binding. Adherence to defined protocols allows runtime binding of independent processes.

Patterns A pattern is a common abstract solution to a common abstract problem that Can be tailored to a given situation Has predefined characteristics Abstraction level of patterns Business Analysis Architecture Design Implementation (Idioms) Test Patterns (or guideline to testing patterns)

Relationship of Tactics to Patterns An architect usually chooses a pattern or a collection of patterns designed to realize one or more tactics. However, each pattern implements multiple tactics, whether desired or not.

Famous Pattern (Style) categories Data-centered Repository Blackboard (publisher-subscriber) Structural solution to integrability of data Scalability Modifiability Client Client Shared Data Client Client

Famous Pattern (Style) categories Dataflow Bach sequential Pipes and filters Reusability Modifiability Not interactive Poor performance

Famous Pattern (Style) categories Virtual Machine Interpreter (e.g. Adaptive Object Model) … Portability Simulation Adaptability Low performance

Famous Pattern (Style) categories Call and Return Main program and sub-routine Modifiability Remote procedure call Performance tuning Object-oriented or abstract data type Reuse Layered Portability

Famous Pattern (Style) categories Independent components Communicating Processes Scalability Event Systems Modifiability

Example: ATM Software Develop 3 different architectures for ATM software and compare them regarding fulfillment of quality attributes. ATM = Automatic Teller Machine User operations: Insert card and enter PIN Withdraw money Check Balance

Shared-Memory Style

Abstract Data Type Style

Layered Style

Analysis and Comparison Shared-Mem ADT Layered Performance 3 2 1 Change account record format New service: close account and withdraw the remained balance Portability Availability and Reliability Buildability and Integrability Sum 9 13 15

Software Architecture Analysis Method (SAAM)

Topics about SAAM Why analysis? Selecting system to buy Developing different architecture and comparing them What specifies the metric for evaluations? Goal(s) How to evaluate SA with regards to goals? Write scenarios to analyze the system against goals

Steps of SAAM Develop Scenarios Describe Candidate Architectures Classify Scenarios Direct Indirect Perform Scenario Evaluations (for indirect scenarios) Reveal Scenario Interaction

Scenarios for ATM example 0. Withdraw money New service: close account and withdraw the remained balance Change hardware Change DBMS Change account record format

ADT Style Architecture

Layered Style Architecture

Comparison Scenario 0 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Contention ADT Direct - Layered + For these scenarios layered architecture is superior to ADT architecture

Other Issues about SA Product-Line Architectures Architecture Description Languages (ADL) Run-Time Re-Configuration of Architecture Architecture Reconstruction Information Architecture Enterprise Architecture

Good References SEI Software Architecture Series: Software Architecture in Practice, 2nd Ed. Documenting Software Architecture Evaluating Software Architecture: Methods and Case Studies Software Product Lines

Good References Pattern Books: Design Patterns: Elements of Reusable Object-Oriented Software Analysis Patterns: Reusable Object Models Pattern-Oriented Software Architecture