Software Architecture Engineering and Management

Models, methods and tools for software architecture management and utilization, extracting high-level architecture models directly from the code, support for manual and automatic architecture analysis, synchronization with enterprise information repositories

Motivation 

Software engineering is an engineering discipline whose focus is the cost-effective development of high-quality software systems. Software architecture takes a central part in software engineering. As an abstraction of a software system it is the main artifact that is used for reasoning about central system qualities like like performance, security, and scalability, to name a few. Early design decisions are usually long lasting and expensive to change later in the development. For developers architecture is a blueprint of the system that guides and constrains its implementation. For stakeholders, software architecture is an important means for communication and documentation.

Architecture-related engineering activities like software architecture evaluation require an extensive amount time and human resources and thus lack the cost-effectiveness as one of the main goals of software engineering. Formalization and tool support contribute to more cost effectiveness, but formalized approaches for architecture description and analysis, like Architecture Description Languages (ADLs) often lack tool support and are not aligned well enough with system implementation. Also formalized approaches often lack multiple views on a software architecture, require specialized knowledge hindering their adoption, and are too rigid to be used for communication.

Goals

The main aim of this project is to bridge the gap between formal and non-formal approaches to software architecture description and analysis based on a central semi-formal architecture model that is explicitly designed to be (continuously) aligned with a system implementation and to integrate both automatically extracted and manually created architecture information. A set of tools aims to support multiple views on the captured information and various kinds of both automatic and manually architecture analysis.

• Develop models for describing and analyzing software architectures
• Integrate software architecture information into the system implementation
• Extract architecture information directly from code; continuously synchronize architecture information with code

Results

The main result is the LISA (Language for Integrated Software Architecture) approach, which has evolved over several years and now has many facets. Some of them are:

• A set of integrated models for describing software architecture on different levels of abstraction and granularity
• Models for representing code structures and relationships (e.g., classes and dependencies), components, services, and connectors, requirements, design decisions, and rationale, and features, variation points, variants, and constraints.
• A set of tools for architecture extraction, synchronization with external resources, architecture visualization, and both manual and automatic architecture analysis
• Extraction of architecture information from a system implementation through different component-model specific parsers. Support for heterogeneous systems through support for different component and service models within one system architecture.
• Support for different kinds of continuous architecture analysis, including a rule editor for high-level specification of reference architecture conformance analysis
• Support for manual analysis and review through different views. Generation of diagrams from the architecture model. Annotation of visualized elements in the review process. Creation of review reports and reviews based on review templates.
• Prototypical integration of AOP concepts like aspects, joinpoints, point cuts, and advices
• Provisioning of architecture models and functionality through a REST-API
• Support of automatic capturing of traces from requirements and design decisions to design and code artifacts
• Application in the context of a Service-Oriented Architecture (SOA) in the banking domain

(please have a look at the related publication list for a more exhaustive list of the provided features and investigated issues)

Duration
since 2008

Partners
GRZ IT Group
Software Competence Center Hagenberg (SCCH)

Subprojects and related projects
Software Architecture Knowledge Management
Variability Management in Software Architecture

Contact Rainer Weinreich

Publications