Abstract
Some of current Information-centric cellular networks strongly rely on clustering techniques to group users with similar preferences or behavior and assign them to a common base station. As a result, the most useful contents are kept near to users, thereby alleviating backhaul bottlenecks. Despite the performance gains, such strategies frequently adopt the same replacement policy, irrespectively the users’ profile. This paper presents a simulation-based study that evaluates the performance of cache replacement policies through clusters formed by users according to their music listening habits. It shows the benefits of using caching strategies according to users’ pattern of behavior on downloading songs. To accomplish such goals, we carried out a data mining based analysis of traces obtained from a music streaming service to identify user groups that share similar listening habits. The resulting clusters supported the experimental study carried out to evaluate the performance of three usual replacement policies.
Stéfani Pires
PhD in Computer Science
My research interests include computer networks, information centric networking.
Adriana Viriato Ribeiro
PhD Student
My research interests include computer networks and information centric networking.
Mateus Sousa
PhD Student
My research interests include information centric networking, vehicular ad hoc networks, and blockchain.
Allan Freitas
Professor of Computer Science
My research interests include Distributed Computing, in particular, Distributed Algorithms and Distributed Systems.
Leobino Sampaio
Professor of Computer Science
My research interests include Information-Centric Networking (ICN), in particular, the NDN Architecture, applied to mobile scenarios, such as VANETs, FANETs, and IoT.