ENERGY EFFICIENT TECHNIQUE AND ALGORITHM BASED ON ARTIFICIAL INTELLIGENCE IN CONTENT DELIVERY NETWORKS

Abstract

Due to the large number of servers and network infrastructure required to deliver content to users, content delivery networks (CDN) consume a large amount of energy. CDN use several strategies to reduce energy consumption, such as server consolidation, dynamic provisioning, and load balancing. However, these strategies do not take into account the popularity of the content being presented. Therefore, a mechanism to improve energy efficiency based on content popularity has been developed in CDN. The main function of the mechanism is to make maximum use of the cache servers' memory capacity at the expense of optimal service to user requests and to increase the service performance of cache servers to user requests. To achieve this, based on machine learning algorithms using user requests and attributes of video files, predicting the probability of video content becoming popular and storing videos with the highest popularity index on edge cache servers is caught.