Optimal Network Coding based on Machine Learning Methods for Collaborative Networks
Javier Mendoza-Almanza, Francisco de Asís López-Fuentes · 2019
Multimedia content distribution over the Internet has increased significantly during the recent years. Almost 50% of the world population has access to the Internet, which may overwhelm its content distribution capacity and saturate communication links. Network coding is a coding method used to increase throughput, scalability or resilience in the communication networks. In this paper is presented a dynamic system with network coding using on machine learning technique for collaborative networks. Our network coding approach is based on XOR logic operations, while collaborative scheme is supported by a Peer-to-Peer (P2P) network. Our proposal uses a coordinator server to synchronize all nodes, and to assign their roles during a network coding operation. Coordinator server also ensures that the entire process is completed. Our results show the advantages of combining network coding with techniques of automatic learning, which are based on the comparison between decision tree and K-means. These techniques help to coordinator server to assign roles to the different servers, in order to guarantee and make efficient the network coding process.