Research On Traffic Engineering Based Admission Control In MPLS Network

With the rapid development of various new network applications and private networks(VPN) catering for their needs, IP network is gradually becoming a unite medium transmitting audios,videos and data, changing from only supporting best-effort services to offering integrated or Quality of Service ( QoS) promise services. Among various new switch techniques, MPLS(Multi-protocol Label Switching)has its unique advantages on supporting QoS,TE(Traffic Engineering) and network's expansibility.Admission control and routing are tow key aspects for transmission network supporting advanced QoS and TE. However, traditional TE solutions usually rely exclusively on routing algorithms to achieve optimization on network resource usage, which usually find an explicit route according to some optimization criterias with constraint based routing (CBR) algorithm, while admission control is simply limited to a local check of resource availability on the explicit route without any optimization purpose, which usually lead to network congestion and reduction of network performance.Basing on the edge admission control frame in [11], this paper use the cost function in [34] which provides the unite IP network's congestion situation measure as the admission threshold. The new LSP eatablishing request will be rejected if it results the cost rise exposively, even though the CBR could find a suitable route for it. The simulation results show that the admission control which is not"greedy"can avoid network congestion in some way as well as increase the network throughput comparing to admission control which is only used to check local resource availability.In the paper, first introduce the MPLS-TE mechanism and analyse its core techniques and advantages. Then, present the existent admission control algorithsms in and abroad. After that, introduce in details the features and meanings of the cost function and the admission control algorithm using it as admission standard. At last, give the estimate of the admission control algorithm using the simulaition software NS.