| Following the development of the network communication and increase of users, the congestion problem has also become more severely. Chengdu Unicom is one of the most important ISP in Sichuan Province, it holds the most professional networking maintenance team with ample experience and can supply the best service to the users around Chengdu city and surroundings. But now Chengdu Unicom also faces several problems such as lack of new business and plenty of complains from users.Network Congestion may lead to wholly deterioration of network performance,which includes increase of packet loss rate, enhancement of peer to peer delay, even the collapse of the whole system. So in order to solve the problem, improve the performance and the network communication quality, we should decrease the probability of congestion appearing. In this paper, we analyze plenty of materials and find out the root of the network congestion, at the same time we execute the new algorithm in key links aiming at network core. Then we get the high QoS promising environment. In this research we have completed jobs as follows:Through the analysis and conclusion about the reason, mechanism and algorithm of network congestion, we proposes the idea of path optimization to solve the problem. Firstly we analyze the effect of path optimization to network route selection, and deeply research the effect of the method. At the same time, based on the analysis of network congestion we discuss the condition and context of network path optimization which can be used to solve the problem finally. At last we complete the optimization plan of Chengdu Unicom core network, in which we emphasize on the design and implementation of network congestion control.Through the network optimization the Chengdu Unicom core network has become reliable network with reasonable structure and stability, it has also changed a lot in network congestion control. In recent half one year, there is no failure causing by structure and tactics design. And the complains from users has also decreased, which proves that the whole optimization plan is useful and safe. |