The proliferation of smart mobile devices has led researchers to consider Opportunistic Networks (OppNets) as a way to support novel content-centric applications, like file sharing, mobile data offloading, collaborative computing, etc. Special attention is given to mobile data offloading, due to the rapid increase in data traffic demand that has overloaded existing cellular networks. In offloading through OppNets, the cellular network distributes popular content only to a few "edge nodes" (e.g. user devices) which opportunistically forward it to any other interested user. A crucial factor for the performance, the design, or even the feasibility of an offloading mechanism (or an opportunistic content-centric application, in general) is the popularity of the distributed contents. E.g., a popular content might be easy to disseminate, whereas it might not be efficient to offload a content with a few nodes interested in it. To this end, we study the effects of content popularity on the performance of opportunistic content distribution and offloading. We derive simple, closed-form expressions that predict performance as a function of content traffic patterns and system parameters. We use our results to answer questions, like: which contents should be offloaded? what is the (monetary) cost incurred for the operator? how to optimize an offloading mechanism?; and provide initial insights into various design trade-offs involved for the operator, like, pricing strategies, network planning and dimensioning, CAPEX vs OPEX costs, etc.
Pavlos Sermpezis is a Postdoctoral Researcher at the Institute of Computer Science (ICS) of FORTH, Greece. He is working with Prof. Xenofontas Dimitropoulos and the INSPIRE group of the Telecommunications and Networks Lab (TNL). From November 2011 to February 2015, he was a PhD candidate at EURECOM, Sophia-Antipolis, France, under the supervision of Prof. Thrasyvoulos Spyropoulos. Before that, he received his Diploma in Electrical and Computer Engineering, from the Aristotle University of Thessaloniki, Greece, in 2011.
His current research is on the Internet's backbone, and, in particular, on inter-domain routing problems, using tools and approaches relating to the areas of: Internet measurements, Software-defined networks (SDN), Network economics. A main part of his research (most of which conducted during his thesis) is on modeling and performance analysis for Mobile and Social Networks (MSNs), which he studies using (mainly) analytic tools, like: stochastic processes, random graph theory, network epidemics, complex networks analysis.
He has participated in the european projects "Netvolution" and "SCAMPI", served as Publicity Chair in the IEEE NetSciCom'14 Workshop (joint with INFOCOM) and Local Arrangement Chair in the PAM'16 Conference, been Teaching assistant in the Master courses "SDN" and "Network Modeling" in the University of Crete and EURECOM, respectively, and was an invited speaker in Graduate schools/workshops on Complex Networks and Information-Centric Networks (ICN).
More information about his work and publications, can be found in his personal webpage: https://sites.google.com/site/pavlossermpezis/
