Separable equilibrium distributions for systems with unreliable wireless links

Performance evaluation of systems including, among others, components based on wireless links plays an important role both in the software and hardware engineering. In both cases the earlier evaluation of performance characteristics has been proved to be a key-factor for a successful design of new systems. However, most of the quantitative analysis methods suffers the problem of not being scalable with respect to the number of system components, due to the state space explosion problem. In this paper we address the problem of the quantitative modelling of system with some unreliable wireless links (i.e., where packet loss and retransmissions must be considered) using a set of results of product-form theory. Product-form models present separable equilibrium distributions of the state probabilities thus making the derivation of the desired performance indices (e.g., mean number of customers or mean response time) computationally tractable. This work exploits recent advances in this field, e.g., the Reversed Compound Agent Theorem, to efficiently evaluate the performance of systems with unreliable wireless links.