Computing Reviews

Optimal online scheduling with arbitrary hard deadlines in multihop communication networks
Mao Z., Koksal C., Shroff N. IEEE/ACM Transactions on Networking24(1):177-189,2016.Type:Article
Date Reviewed: 07/26/16

This carefully written paper from Mao et al. makes important contributions to optimal online packet scheduling in the context of hard deadlines and a multihop wired network. A modification of the familiar earliest deadline first (EDF) algorithm, called the work-conserving earliest deadline first (WC-EDF) algorithm, is helpful in advancing results, and then the new admission control and packet scheduling algorithm leads to even more substantive results. This algorithm has O(n log n) performance, where n is the maximum route length among the packets, and indicates a highly satisfactory efficiency.

The introductory section provides a literature review and outlines the strategy pursued in the rest of the paper. The next two sections provide a formal problem statement and a motivating example to clarify the essence of the problem statement. The fourth and fifth sections form the core of the research in terms of technical detail that includes formal algorithms, propositions, and their proofs (one of which was moved to an appendix), as well as some diagrams that illustrate the working of these techniques. The sixth section shows numerical results from applying the algorithms, and the concluding section effectively summarizes the research presented in the paper.

This paper contributes to better analysis of online versus offline algorithms for multihop settings, and the new methodology for packet scheduling allows more rigorous analysis, as shown in the paper, leading to future applications to flow scheduling and software-defined networks. The paper is an interesting mix of reading levels, from light to highly mathematical.

Reviewer:  M. G. Murphy Review #: CR144630 (1611-0829)

Reproduction in whole or in part without permission is prohibited.   Copyright 2024 ComputingReviews.com™
Terms of Use
| Privacy Policy