Causal modeling of a video-on-demand system using predicate/transition net formalism

Abstract

We give a high-level net model for a specific video on demand system. This system has some very interesting modeling features: FIFO queues, the bunching property, parallel machine scheduling and a complicated resource allocation mechanism. The bunching property is the feasibility of several requests being served together by one resource. The resource allocation mechanism for this video on demand system has a special refusal by choice feature which means that the resource is not necessarily allocated even though it is available. The predicate/transition net formalism has been used in the modeling and reachability analysis in studying the properties of the model. The causal model developed here allows a natural extension for the use of time or stochastic net theoretical methods for performance analysis.