Bandwidth tomography with probabilistic routing
Date
2024
Authors
Zhang, Xiaowei
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The demands for maintaining optimal Quality of Service (QoS) have become higher with the rise of new Internet applications. Such demands include the need to monitor internet performance metrics such as bandwidth. Nevertheless, directly measuring the bandwidth of every individual link is often impractical due to the sig- nificant overhead involved in the measurement. Instead, it is essential to estimate the bandwidth of individual links using the measurements from only a few network nodes, a process known as bandwidth tomography. A recent work by Feng et al. introduced an algorithm to address this challenge, which determines the error bounds for link bandwidth estimates. This work, however, assumes deterministic routing. In real-world scenarios, routing may be probabilistic rather than deterministic. In this thesis, we extend the applicability of this existing work from the deterministic routing scenario to the probabilistic routing scenario by developing an algorithm that maps end-to-end available bandwidth values to their corresponding paths with high prob- ability. Such an extension pushes the state-of-the-art bandwidth tomography toward a more complex scenario.
Description
Keywords
Bandwidth, Tomography