Buyukates and Bastopcu win Best Student Paper Award at IEEE SPAWC 2021

A paper written by two ECE Ph.D. students and their advisor has won the Best Student Paper Award at the 2021 IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC 2021). “Age of Gossip in Networks with Community Structure” was written by ECE Ph.D. student Baturalp Buyukates; recent alum Melih Bastopcu (ECE Ph.D. 2021), now a postdoctoral researcher at the University of Illinois Urbana-Champaign; and their advisor, Professor Sennur Ulukus (ECE/ISR). Both Buyukates and Bastopcu also were honored recently with the Institute for Systems Research's 2021 George Harhalakis Outstanding Systems Engineering Graduate Student Award.

About the paper

Introduced to quantify timeliness in real-time status updating systems, the age of information metric has received significant attention across information, communication, networking, and queueing theory fields. Recently, a new age metric—version age—has appeared in the literature. Considering each update at the source as a version change, the version age counts how many versions out-of-date information is at a particular receiver, compared to the version at the source.

In this paper, the authors investigate version age scaling in general gossip network models which exhibit a community structure. They consider a network consisting of a single source and n receiver nodes that are grouped into m equal size communities, i.e., clusters, where each cluster includes k nodes and is served by a dedicated cluster head. The source node keeps versions of an observed process and updates each cluster through the associated cluster head. Nodes within each cluster are connected to each other according to a given network topology. Based on this topology, each node relays its current update to its neighboring nodes by local gossiping. The authors use the version age metric to quantify information timelines at the receiver nodes.

Disconnected, ring, and fully connected network topologies are considered for each cluster. For each of these network topologies, the authors characterize the average version age at each node and find the version age scaling as a function of the network size n. Their results indicate that per node version age scalings of O(√n), O(n^(1/3)), and O(log n) are achievable in disconnected, ring, and fully connected cluster models, respectively. Finally, through numerical evaluations, the authors determine the version age-optimum (m, k) pairs as a function of the source, cluster head, and node update rates.

Published October 6, 2021