Distribution-Based Energy Efficiency Analysis of Intelligent Content Service Network

Yong Tan,Xiaolong Fu,Xiaofeng Zhong,Dongxing Jiang

DOI: https://doi.org/10.1109/vtcfall.2016.7880965

2016-01-01

Abstract:Wireless communication network is always resource limited, which is hard to satisfy the rapidly growing QoS requirements of users, such as high data rate and low energy consumption. The simple and straightforward solution is to alleviate the network resources and energy consumption. This paper proposes Intelligent Content Service Network (ICSN) to satisfy users' requirements based on the concept of spatial-temporal decoupling. The main idea of ICSN is to make contents predictions by analyzing users' accessing history, then choose the right time and location to accomplish content transmission according to the predicting results. Experiments are conducted under two typical and ubiquitous scenarios, which are Tsinghua XuetangX and THU Online Learning Platform, in order to illustrate the energy efficiency of ICSN, here represented by the network traffic flow reduction rate. After analyzing the users' behaviour, we put forward predicting methods according to their respective characteristics, and results show that the network traffic is reduced by 23% and 68% under these two scenarios respectively. Moreover, for ideal condition that both of time and location can be predicted precisely, the reduction rate reaches 98%. The results show that ICSN can greatly improve the energy efficiency, which depends on the prediction precision of content transmission time and location.