A Survey on Process Algebraic Stochastic Modelling of Large Distributed Systems for Its Performance Analysis

Abstract

The modern distributed systems have not only functional requirements (i.e. Absence of deadlock, live lock etc.) but also have non-functional requirements (i.e. Security, reliability, performance, Quality of Service (QoS) etc.). The methods for checking their correctness and analyze their performance is at very primitive stage. In the last few decades, formal verification techniques such as process algebras offer a powerful and rigorous approach for establishing the correctness of computer systems. The classical process algebras were concerned with the functional aspects of the concurrent systems. Stochastic extension of process algebras is developed to add quantification to process algebra models. In this paper, we present a comprehensive survey on different stochastic process algebras to find the recent trends of development and limitations of the various developments of stochastic process algebras. We also discuss state space explosion problem which is the major problem in classical stochastic process algebras and different approaches described by the different authors to tackle with the state space explosion. Finally we suggest the possibility to use, one of the methods to tackle state space explosion problem called fluid flow approximation, along with stochastic process algebras to model and do performance analysis of routing calculi, which are extensions of distributed calculi. These routing calculi are developed to model a distributed network with the intention to demonstrate the cost of communication between the communicating processes.