Indian Journal of Science and Technology

Abstract

The evaluation and performance analysis of software architecture at the design level increases the quality of the software and also reduces the cost of rework during the later stages of the product. The derivation of performance results of a software product, during the early stages of the software life cycle can be achieved by quantitatively evaluating the software performance model. There has been lot of research identifying the methods of evaluating software (Booch, 2001). The evaluation process starts by analyzing the performance model which is derived from the software model annotated with suitable usage profiles. This paper provides a simple approach to convert the software models represented as Unified Modeling Language (UML) diagrams using the profile for Schedulability, Performance and Time specifications (SPT) into Layered Queuing Network (LQN) performance models. The paper mainly illustrates the conversion process from UML to LQN, and also substantiates the method by a simple example.
The evaluation and performance analysis of software architecture at the design level increases the quality of the software and also reduces the cost of rework during the later stages of the product. The derivation of performance results of a software product, during the early stages of the software life cycle can be achieved by quantitatively evaluating the software performance model. There has been lot of research identifying the methods of evaluating software (Booch, 2001). The evaluation process starts by analyzing the performance model which is derived from the software model annotated with suitable usage profiles. This paper provides a simple approach to convert the software models represented as Unified Modeling Language (UML) diagrams using the profile for Schedulability, Performance and Time specifications (SPT) into Layered Queuing Network (LQN) performance models. The paper mainly illustrates the conversion process from UML to LQN, and also substantiates the method by a simple example. Keywords: Usage profiles, Unified modeling language, Performance model, Model Driven DevelopmentUsage profiles, Unified modeling language, Performance model, Model Driven Development