Congestion Control In The Red Algorithm Improvements

With the rapid development of Internet, various applications based on Internet have emerged and the data flows on Internet are increased abruptly. Network congestion has become a serious problem and congestion control has been the focus of attention in the field of network.Currently, there are two kinds of the schemes of network's congestion control. The first is TCP congestion control strategies, the second is IP congestion control strategies. In this paper, we have deeply researched the two types of congestion control strategy.We study TCP congestion control process including slow start,congestion avoidance,fast recovery and fast retransmit ,and then we introduce the congestion control schemes of TCP Tahoe, Reno,Veagas and Sack, and compare between the capability of the representational Vegas and Reno algorithm, and make an experiments based on NS.We introduce the fundamental of IP congestion control strategies,Queue Management and Queue scheduling, and then we study the fundamental of RED,introduce some importanct amelioration RED algorithms.We introduce a new algorithm to impove RED capability . We ameliorate RED algorithms based on self-similarity of the network data flow and unfair competition characteristics of RED.The system can judge the trend of the data flow using network bandwidth based on packets classes and the packets speed. So,the system can calculate the probability of droping packets.The new RED algorithms has been proved it has advance fair capability by our experiments.