TCP Congestion Control Algorithm for Heterogeneous Networks

The end-to-end bandwidth, the Round Trip Time (RTT) and the packet loss rate vary over many orders of magnitude in Internet, which cause significant performance degradation of Transmission Control Protocol (TCP). To cope with this, a path-adaptive TCP congestion control protocol, called INVerse Sigmoid (INVS), is proposed. INVS employs an exponential function based convex growth function of congestion window to improve the path utilization. An adaptive increase factor in the growth function is introduced to match the window growth rate with the path condition. An adaptive queue-threshold in the loss distinguishing scheme is used to enhance the performance of TCP over wireless links. The performance analysis and evaluation show that INVS improves the throughput, the fairness, the total link utilization and the RTT-fairness of TCP congestion control algorithm.