Architecture-based approach to support the evaluation of alternative design decisions

Component-Based System (CBS) is a promised approach to build applications from deployed components. It provides efficiency, reliability, maintainability. Interpreting the results of performance analysis and generating an alternative design to build system from component (Hardware/Software) is a great challenge in the software performance domain. There are so many options to compose the system. Span of design space hinders the selection of the appropriate design alternative. Currently, Meta-heuristics such as Genetic Algorithm (GA) methods are used to solve the problem. In recent investigations Particle Swarm Optimization (PSO), an alternative search technique, often outperforms GA when applied to various problems. In this paper, we describe performance prediction approach based on PSO for component-Based system development. PSO technique can be used to effectively generate alternatives design options in spanned design space and facilitate the design decision during the development process. The proposed approach aids developers to effectively trades-off between architectural designs alternatives, because it covers and generates all possible options and provides the best solution. To the best of our knowledge, we are the first who use PSO in software performance prediction, particularly in the context of CBS. To this end, outlines and an example are presented in the paper to describe the approach.