Social Relation Based Cache Distribution Policy in Wireless Mobile Networks

Recent advances in intelligent mobile terminal result in the rapid development of wireless mobile networks, which has inherently properties of high access time, low bandwidth and other shortcoming that pull down the Quality of Service (QoS). Cache technologies are considered as the most favorable tools for dealing with the problems of wireless traffic flow, having attracted tremendous interest from both the engineering and academic fields. However, most of the algorithms and tools developed are essentially based on descriptive models instead of predictive models, which are not capable of processing content traffic in the optimal way for wireless users with strong social characteristics. In this paper, a novel social relation based cache distribution policy is proposed, by which not only interaction among network users is used to detect the influential users among them, but also inherent friend relationships are used to predict desirable contents. These features are utilized to assist cache distribution among caching servers in advance for the purpose of serving the future needs of wireless mobile users. Evaluation analyzes the realistic network dataset, and the simulation results demonstrate the superior performance of the proposed optimization framework over existing systems