Simplifying Wireless Social Caching via Network Coding.

Social groups open up the opportunity for a new form of caching. This paper investigates how a social group of users can jointly optimize bandwidth usage, by each caching network-coded parts of the data demand, and then opportunistically share these parts among themselves upon meeting. First, the problem is formulated as a linear program (LP) with exponential complexity in the number of users. Then, a heuristic algorithm is proposed, which is inspired by the bipartite set-cover problem and operates in polynomial time. For some scenarios, a worst-case performance guarantee of the heuristic with respect to the optimal LP solution is proved. Finally, the performance of the algorithm is assessed using real-world mobility traces synthesized using the SWIM model and from the MIT Reality Mining project data set. The proposed heuristic offers bandwidth savings up to 65% for a waiting time of 30 minutes and up to 28% performance gains with respect to the alternative solutions. These benefits make the algorithm a feasible candidate solution for bandwidth savings. [ABSTRACT FROM AUTHOR]