Priyanka M. Gawande, A. A. Nikose
Abstract: In this paper, Prediction based TRE System to reduce Cloud Bandwidth & Cost is presented which is a receiver based end-to-end traffic redundancy elimination TRE system. This is designed for the cloud computing customers. The Cloud-based TRE system needs to apply a judicious use of cloud resources to combine the bandwidth cost reduction with the additional cost of TRE computation and the storage would be optimized. The main advantage and objective of the Predictive Acknowledgement (PACK) is its capability of offloading the cloud-server TRE effort to end clients. This results in minimizing the processing costs which is induced by the TRE algorithm. Here, the PACK does not require the server to continuously maintain the clients status unlike the previous solutions. This makes suitable for pervasive computation environments that combines the client mobility and server migration to maintain the cloud elasticity. The TRE based Cost Reduction system is based on a prediction TRE technique, that allows the client to use the newly received chunks for identifying the previously received chunk chains, which in turn can be used as the reliable predictors to future transmitted chunks. Thus we present a fully functional PACK implementation, which is transparent to all TCP-based applications and network devices. Finally, we analyze the Prediction Acknowledgement benefits for cloud users, using the traffic traces from various sources. PACK, TRE, cloud-server, client mobility, TCP.
Keywords: PACK, TRE, cloud-server, client mobility, TCP