Product-form solutions for integrated services packet networks and cloud computing systems

TL;DR

This paper derives product-form solutions for the stationary distributions of priority multiclass queueing networks with multi-server stations, aiding performance analysis of large-scale cloud computing systems.

Contribution

It introduces an iterative method to obtain product-form solutions for complex queueing networks, enhancing analysis accuracy for cloud and multi-processor systems.

Findings

Effective algorithm for product-form solutions

Numerical validation against Brownian models

Applicable to large-scale cloud computing networks

Abstract

We iteratively derive the product-form solutions of stationary distributions of priority multiclass queueing networks with multi-sever stations. The networks are Markovian with exponential interarrival and service time distributions. These solutions can be used to conduct performance analysis or as comparison criteria for approximation and simulation studies of large scale networks with multi-processor shared-memory switches and cloud computing systems with parallel-server stations. Numerical comparisons with existing Brownian approximating model are provided to indicate the effectiveness of our algorithm.