Performance Evaluation of Exponential TCP/IP Congestion Control Algorithm

This paper introduces and analyses a class of non-linear congestion control algorithms called exponential congestion control algorithm. It generalizes the AIMD algorithms used for the TCP connections. This algorithm provides additive increase using the inverse of an exponential of the current window size and provides multiplicative decrease using the exponential of the current window size. It is further parameterized by alpha and beta. There are infinite numbers of TCP-compatible exponential algorithms by assuming exponent of different order. This paper analyses the performance of the generalized algorithm (TCP/EXP), for the TCP networks. TCP compatibility of this algorithm is evaluated using the simulations of the implementation of the proposed model. Simulations are done using ns2, a discrete event simulator. The results of simulation are compared with that of the TCP variants such as TCP/Tahoe, TCP/Reno, TCP/New Reno, MIMD and PIPD. The Comparison shows that the proposed algorithm performs better in terms of throughput.