THE EFFICIENCY AND RELIABILITY ANALYSIS OF COMPUTER SYSTEMS BASED ON STOCHASTIC PETRI NETS

The application of one of the most popular extensions of Petri nets - stochastic Petri nets (SSP) for the analysis of the efficiency and reliability of hardware and software of computing systems is considered. An isomorphism between SSP and Markov chains with continuous time is noted, which is the basis of the quantitative analysis of SSP by means of Markov chains analysis tools. The approaches to the construction of operational models of software and hardware of computer system are described. Particularly the specifics of behavioral and algorithmic approaches and the features of setting the absolute and relative load in the SSP-models of programs are considered. A methodology for the analysis of SSP based on the apparatus of Markov chains is presented and examples of constructing SSP-models and calculation of the characteristics of the efficiency and reliability of computer systems based on models of SSP are considered. For the fault-tolerant computer system model, the average expected system performance in stationary mode is estimated. The necessity of using a hierarchical approach to the construction of SSP models due to the high computational complexity of calculating the characteristics of aircraft based on them is noted.

Authors: V. A. Kirianchikov

Direction: Informatics, Computer Technologies And Control

Keywords: Efficiency, reliability, stochastic Petri net, resource consumption parameter, operational SSP-model, set of reachable states, continuous time Markov chain, ergodic Markov chain, transition intensity, final probability vector

---

View full article