Integration of Reliability and Performance Analyses for Active Network Services

Increasing the quality of software for new telecommunication services requires the joint use of different testing techniques. For instance, automatic verification and performance evaluation are necessary to ensure desired throughput and reliability. However, both kinds of analysis were traditionally performed without sharing a common description of the system, and much work and time was wasted constructing different specifications oriented to particular tools. In recent years a lot of research has been carried out to design languages and tools to manage both functional and performance analysis with only one description, specially within communities devoted to process algebras and Petri nets. These homogeneous frameworks remove the inconveniences of maintaining a set of specifications for the same software. In the paper, we explore an alternative approach to keep the specification of different aspects to be analyzed consistent. Taking into account the number and quality of existing tools for modelling and analyzing telecommunication software, we explore semi-automatic methods to integrate these tools in a way that is as transparent as possible for users. Ideally, the designer of new services will provides only one description of the software with its most familiar language, and a toolset will generate particular specifications to analyze each aspect of interest (currently, reliability and performance). Our proposal takes advantages of recent work on model-driven architecture (mda) and xml for automating tool construction. Its applicability is shown in the context of developing new services with the active network paradigm, integrating the features of the model checker spin and the network simulator ns2.