A resequencing model for high speed networks

Abstract

In this paper, we propose a framework to study the resequencing mechanism in high speed networks. This framework allows us to estimate the packet resequencing delay, the total packet delay, and the resequencing buffer occupancy distributions when data traffic is dispersed on multiple disjoint paths. In contrast to most of the existing work, the estimation of the end-to-end path delay distribution is decoupled from the queueing model for resequencing. This leads to a simple yet general model, which can be used with other measurement-based tools for estimating the end-to-end path delay distribution to find an optimal split of traffic. We consider a multiple-node M/M/1 tandem network as a path model. When end-to-end path delays are Gaussian distributed, our results show that the packet resequencing delay, the total packet delay, and the resequencing buffer occupancy drop when the traffic is spread over a larger number of homogeneous paths, although the network performance improvement quickly saturates when the number of paths used increases. We find that the number of paths used in multipath routing should be small, say up to three. Besides, an optimal split of traffic occurs at paths with equal loads.