Performance Analysis of a Video on Demand System Using Stochastic Petri Nets, ACase Study

摘要

In this report the authors study the performance of an adaptive video on demand(VOD) system using stochastic Petri nets. In this system channel resources are allocated more efficiently than in well-known full VOD or near VOD systems. In adaptive VOD a video server does not notify users immediately it receives their requests, but after a time delay (delayed notification). This enables the server to accept more requests, because it can allocate one channel for several users (bunching property). The efficiency can be improved even more by using a proper rejecting strategy, where some requests are rejected, although there are channel resources available (refusal by choice). The authors give a system model for this adaptive VOD system using stochastic Petri nets. Discrete event simulation is used to determine the performance of the system; numerical performance indicators are computed using the software toolkit TimeNET. (Copyright (c) 1996.)