How to evaluate/measure a software architecture?
Session presented at iLean, 23/3/2011Learning Objectives:
After this session, participants will have a better understanding of:
- What quality attributes such as reliability, performance, modifiability, etc. mean, and how to specify them in a measurable way.
- How quality attributes affect a software architecture
- Why architecture analysis is important
- How QAW and ATAM are performed
- Do’s and don’ts when performing architecture evaluations
Abstract
Most complex software systems should be modifiable, performant, secure, flexible, portable, reliable, etc. But how do you know whether the software architecture for a system will achieve these goals (or not) without having to build the system first? Is there a structured way to analyze/measure a software architecture?
In this interactive session, we discuss to which extent architectures can be evaluated and we highlight approaches that support architecture evaluation in a systematic way, including the Quality Attribute Workshop (QAW) and the Architecture Tradeoff Analysis Method (ATAM). The QAW is a lightweight, stakeholder-oriented workshop aimed at prioritizing and specifying architecture qualities in a measurable way. The ATAM is an established method for systematically evaluating software architectures for fitness of purpose. The ATAM exposes architectural risks that potentially inhibit the achievement of quality goals and the system’s business/mission goals. Empirical measurements show that a structured architecture evaluation saves about 10% of the total cost of a software project.