As Editor-in-Chief of The Computer Journal, published by Oxford University Press on behalf of BCS, The Chartered Institute for IT, it is my honour to select the recipient of the annual Wilkes Award, awarded to the authors of the best paper published in the previous year’s volume. The key judging criteria for the Award are originality and quality of theme and treatment, as assessed by an expert judging panel.
We are pleased to announce this year’s winning, runner-up and shortlisted papers for the 2021 Wilkes Award for Volume 63 (2020):
Winning paper
RIoT: A Routing Protocol for the Internet of Things
Muhammad Omer Farooq
The Computer Journal, Volume 63, Issue 6 (June 2020), pp.958-973
https://doi.org/10.1093/comjnl/bxaa012
This paper addresses a recognised weakness in RPL, the standard routing protocol for low-power and lossy networks such as the Internet of Things: its control message overhead leads to poor performance for point-to-point and point-to-multipoint communication, particularly in its non-storing mode. The author proposes RIoT, a routing protocol supporting the same three communication patterns — multipoint-to-point, point-to-point and point-to-multipoint — but constructing the harder point-to-point and point-to-multipoint routes with substantially lower control overhead. RIoT also introduces a multi-gateway architecture that reduces the memory each node needs to store a forwarding table, and extends to mobile devices, a use case the standard protocol does not support. Implemented in the Contiki operating system and evaluated through both emulation and real testbed experiments, RIoT showed statistically significant improvements over RPL in packet delivery ratio (up to 30% higher, depending on communication type), end-to-end delay and control overhead, along with better resilience to network partitioning in mobile scenarios. A clearly motivated protocol design, meaningful extensions over the established standard, and thorough empirical validation made this a strong choice for the inaugural Wilkes Award of my editorship.
Runner-up
Hidden Markov Model-based Load Balancing in Data Center Networks
Binjie He, Dong Zhang and Chang Zhao
The Computer Journal, Volume 63, Issue 10 (October 2020), pp.1449-1462
https://doi.org/10.1093/comjnl/bxz142
The runner-up paper addresses path selection across the multiple parallel routes that modern data centre networks provide. Centralised load-balancing algorithms generally make better path choices than distributed ones, but only by monitoring bandwidth on every link in the topology, which becomes costly at scale. The authors recast path selection as a hidden Markov model, trained on data generated by a conventional centralised algorithm, so that the resulting model needs to monitor only a fraction of the network’s links to make comparable decisions. Tested by simulation across fat-tree and VL2 data centre topologies, the approach roughly halved path-selection time relative to existing methods while achieving throughput close to that of the traditional approach — a different route to the same goal of efficient data centre routing.
You can also explore the shortlist of papers that were selected by the Section Editors for the judging panel’s consideration; all papers are freely available. You can also browse previous winners of the Wilkes Award. Also, thank you to the Wilkes Award judging panel:
- Professor Steve Furnell, University of Plymouth, UK
- Professor Fairouz Kamareddine, Heriot-Watt University, UK
- Professor John Shawe-Taylor, University College London, UK
The award is named after Sir Maurice Wilkes, who was Director of the Cambridge Computer Laboratory throughout the development of stored program computers starting with EDSAC; inventor of labels, macros and microprogramming; and with David Wheeler and Stanley Gill, the inventor of a programming system based on subroutines.
Digital Society & Policy 