Publication Type : Journal Article
Publisher : arXiv:2002.02522v1[cc.NI]
Source : arXiv:2002.02522v1[cc.NI], Feb 6, 2020
Url : https://arxiv.org/abs/2002.02522
Campus : Amritapuri
School : School of Computing
Year : 2020
Abstract : This work addresses the problem of evaluating optimal link capacities of a packet-flow network for the objective of congestion minimization. We present a simple model of packet flow in networks and present a numerical approach to evaluate packet flow probability mass function at any arbitrary edge of the network for a given routing algorithm and traffic rate. We further discuss techniques of assigning optimal capacity at each edge for attaining desired minimized congestion and discuss related trade-offs. Our framework is built around the assumption of Poisson traffic, however the numerical approach fits for any general distribution of packet influx. Lastly, we define metrics of global performance of link capacities allocation and discuss the effect of network structure on capacity allocation and performance.
Cite this Research Publication : S. Pal, A. Chatterjee, D. Bakshi and A. Mukherjee, “Link Capacity Distributions and Optimal Capacities for Competent Network Performance”, arXiv:2002.02522v1[cc.NI], Feb 6, 2020