Deep learning based predictive analytics for decentralized content caching in hierarchical edge networks

Abstract

Content centric network is a state-of-the-art networking architecture for content distribution and content caching. However, it is inefficient to cache every content in each network device. The modern edge computing technology opens the door for content caching in the edge of the network. However, still we have to decide which contents we should cache and which content we should replace from the cache. Deep learning based predictive analytics can play an important role in selecting contents for caching purposes. In this research, we will use Long shortterm memory(LSTM) based Recurrent Neural Network(RNN) for decentralized content caching at the hierarchical edge of the network.