Contact duration-aware cooperative cache placement using genetic algorithm for mobile edge networks

Abstract

Caching popular content at the base stations cooperatively is an effective solution to reduce the user-perceived latency and overwhelming data traffic by bringing content close to the user in a cellular network-based Mobile Edge Computing (MEC) architecture. Most of the existing literature assumes static network models where all the users remain static throughout the data transfer time, and the user can download the requested content from the associated base station. Caching content by considering user mobility and randomness of contact duration is an important issue which has been addressed in this work. We consider the cache placement problem in a realistic scenario where users move at different speeds. The moving users that are connected to the multiple base stations intermittently may not download full content because of contact duration. This, in turn, increases the overall delay in downloading the content for mobile users. The cache placement problem is formulated as mixed-integer nonlinear programming to maximize the saved delay with capacity constraint. The user mobility and contact duration are modeled with a Markov renewal process. Further, a greedy algorithm is presented to solve the problem by adopting submodular optimization. For real scenarios that scale to large library sizes, taking into account the computational time, we have proposed a genetic algorithm-based heuristic search mechanism. Extensive simulation results show that the proposed contact duration aware caching scheme significantly improves the performance in terms of hit ratio and acceleration ratio in a real-world scenario as compared with three existing caching mechanisms.