Context-aware distributed multimedia provisioning based on anycast model towards Future Media Internet
Résumé
The Future Internet is foreseen to fully handle multimedia services which have experienced an explosive growth in the recent years. Therefore, improving the scalability of these services is getting crucial. One common approach for this purpose is the deployment of multiple replicas of a server throughout the network. The resulting issue is to direct the user request to the best server. Methods based on application - layer anycast service constitute good means to (1) support the server's replication strategy and (2) process an appropriate server selection algorithm for the client request according to different metrics. This paper presents an efficient context-aware anycast multimedia provisioning scheme. In addition to the clients and servers nodes, the adopted architecture involves enhanced Service Registry (SR) nodes, in a distributed mode. SRs have two main roles: first, they collect and maintain the servers' contexts and their content description; second, they perform the mapping of the anycast address of the client request to the unicast address of the most convenient server based on both the client and servers contexts, including the underlying network conditions. The paper presents the evaluation of the proposed video provisioning scheme through intensive simulations. The promising results led to the launch of real developments to be undertaken in the framework of a large-scale European project ALICANTE which objective focuses on Future Media Internet.