Content Delivery Networks - Q-Learning Approach for Optimization of the Network Cost and the Cache Hit Ratio

Abstract

With an increasing demand for web content delivery, it is necessary to optimize the CAPEX and OPEX costs of the Content Delivery Networks. Ideally, all web content to be requested should be stored in local cache nodes at all times. However, the content demand varies across space and time. In this paper, we propose a Content Delivery Network model that allows us to choose the best trade-off between costs and cache hit ratio.