Comparing the QoS of Internet audio mechanisms via formal methods
ACM Transactions on Modeling and Computer Simulation (TOMACS)
| Hi-index | 0.00 |
Stochastically timed process algebras are emerging as a promising formalism to support performance modeling and evaluation of concurrent and distributed systems. To become well accepted, such formalisms have to be proved to be able to provide readable, modular, scalable formal descriptions of more and more complex systems which can be automatically analyzed with a reasonable computational effort. Because of its complexity, the ATM technology could be used as a valid benchmark to test the adequacy of the stochastically timed process algebra approach. To this aim, in this paper we consider an ATM switch which supports the four service categories (CBR, VBR, ABR, UBR) defined by the ATM Forum and implements explicit rate marking for ABR traffic, backward explicit congestion notification, and connection admission control. We show that this complex system can be modeled in a compositional and scalable way with the stochastically timed process algebra EMPA. Moreover, we resort to the EMPA based software tool TwoTowers to derive some performance measures of interest for the ATM switch, such as the cell loss ratio, the queue length distribution, and the link utilization for the outgoing link in several different configurations.