The authors present a framework in which the performance analysis of systems is driven by the semantics of their specification. They use an enhanced structural operational semantics based on the higher order pi-calculus. Rates are assigned to the enhanced labels, encoding their derivation. The rates reflect the possible architecture on which the applications may run. Afterwards, a continuous-time Markov chain is associated with the transition system of a given process. From its stationary distribution, the performance of the process is evaluated.

The authors illustrate their approach with several examples of increasing complexity. These examples are used to compare their approach with case studies known from literature. The framework presented is quite interesting, and seems to be a good way to include performance constraints in the specification and implementation of a system. It is claimed that this approach can be scaled up to real-world applications. It will be interesting to see if the authors are able to prove this. The paper is well written, and the authors give many explanations that help the reader to understand the framework presented. Nevertheless, some background on process calculi and/or Markov chains would have been helpful.