Adaptive distributed compression technique utilizing intermediate network nodes

Abstract

Data networks are experiencing an unprecedented growth in traffic resulting in a multitude of challenges including high network congestion, latency, and packet loss rates. These challenges greatly affect the performance of the network and have a drastic impact on the Quality of Service (QoS) for various applications. In order to reduce network congestion and in turn, maintain the desired QoS, an adaptive and distributed compression/decompression scheme is proposed. The proposed scheme performs the compression and decompression processes within the queues of the intermediate routing nodes of a network. The payload of a packet is compressed based on the packet queuing time. The average expected queuing time of packets is determined instantaneously within the intermediate nodes of a network by employing the Pollaczek Khinchin (PK) equation. Using Network Simulator 3 (NS3), the proposed scheme was simulated under different network conditions and compared with non-compressed transmission sessions of different flow rates. The simulation results show a dramatic improvement in network performance. More specifically, the number of lost packets and one way latency are reduced by at least 22.50% and 18%, respectively.