Download now Free registration required
The existence of significant amount of correlation in the network traffic has stimulated the development of in-network traffic reduction techniques since end-to-end universal compression solutions would not perform well over Internet packets due to the finite-length nature of data. Recently, the authors proposed a memory-assisted universal compression technique that holds a significant promise for reducing the amount of traffic in the networks. The idea is based on the observation that if a finite-length sequence from a server (source) is to be compressed and transmitted over the network, the associated universal code entails a substantial overhead.
- Format: PDF
- Size: 122.04 KB