Definition and validation of design metrics for distributed applications

As distributed technologies become more widely used, the need for assessing the quality of distributed applications correspondingly increases. Despite the rich body of research and practice in developing quality measures for centralised applications, there has been little emphasis on measures for distributed software. The need to understand the complex structure and behaviour of distributed applications suggests a shift in interest from traditional centralised measures to the distributed arena. We tackles the problem of evaluating quality attributes of distributed applications using software measures. Firstly, we present a measures suite to quantify internal attributes of design at an early development phase, embracing structural and behavioural aspects. The proposed measures are obtained from formal models derived from intuitive models of the problem domain. Secondly, since theoretical validation of software measures provides supporting evidence as to whether a measure really captures the internal attributes they purport to measure, we consider this validation as a necessary step before empirical validation takes place. Therefore, these measures are here theoretically validated following a framework proposed in the literature.