In recent advancement in cloud computing, Traffic Redundancy Elimination (TRE) offers an effective solution to reduce the bandwidth cost. It is found that both short-term and long term data redundancy tends to appear in the network traffic and the TRE – trace driven approach captures both the data traffic redundancy. In this paper, we hence improve the design of a cooperative end-to-end TRE solution in order to improve the process of detection and removal of data redundancy between multiple layers, where the operations between them is in cooperative manner. The proposed method uses a self-adaptive prediction algorithm to increase the efficiency of TRE in multi-layer design that uses hit ratio of predictions to adjust dynamically the prediction window size. The experiment evaluation shows that proposed method reduces the operational cost in terms of reduced energy and makespan through proper scheduling.
Volume 11 | 04-Special Issue
Pages: 458-465