Link Delay Estimation via Expander Graphs

TL;DR

This paper explores the use of expander graphs and compressed sensing for network delay estimation, showing that relaxing expansion criteria improves the identifiability of network delays in more topologies.

Contribution

It introduces a relaxed expansion criterion that increases the number of networks where delay estimation is feasible by 30%, advancing network tomography methods.

Findings

Delay estimation is feasible for 30% more networks with relaxed criteria.

Simulation shows accurate delay and congestion detection using l1 minimization.

Most network topologies are not expanders under existing criteria.

Abstract

One of the purposes of network tomography is to infer the status of parameters (e.g., delay) for the links inside a network through end-to-end probing between (external) boundary nodes along predetermined routes. In this work, we apply concepts from compressed sensing and expander graphs to the delay estimation problem. We first show that a relative majority of network topologies are not expanders for existing expansion criteria. Motivated by this challenge, we then relax such criteria, enabling us to acquire simulation evidence that link delays can be estimated for 30% more networks. That is, our relaxation expands the list of identifiable networks with bounded estimation error by 30%. We conduct a simulation performance analysis of delay estimation and congestion detection on the basis of l1 minimization, demonstrating that accurate estimation is feasible for an increasing proportion of networks.