An Adaptive Slow Start Algorithm of TCP Vegas for Large BDP Networks
GU Ming, ZHANG Jun, SU Dong-lin(School of Electronics and Information Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100083, China)
TCP Vegas performs better than TCP Reno with respect to network utilization and stability. But TCP Vegas is inefficient in the large BDP networks. TCP Vegas stops slow start too early, and the congestion window is too small to fully make use of available bandwidth. An adaptive slow start (ASS) algorithm is presented based on the reason and condition of the termination of Vegas slow start. The temporary queue delay is estimated and eliminated before Vegas calculates and judges. Simulation results show that ASS algorithm can get right congestion window and improve the performance.