Comparative analysis of path computation techniques for MPLS traffic engineering

We consider the problem of computing traffic engineered paths for requests having bandwidth and delay requirements, when these requests arrive in the network independent of one another. Providing bandwidth guarantees to applications has been important in networks offering service differentiation. With the increase in the number of real-time applications in the Internet, provision of delay guarantees is also receiving much attention. This necessitates the development of sophisticated path selection algorithms which deviate from the shortest-path routing philosophy in traditional IP networks. While these algorithms perform well from the perspective of satisfying application requirements, they often do not take into account long term effects on the network state. One of the major concerns of a service provider is to run the network at maximum utilization while reducing network costs and preventing congestion in the network. For this reason, providers are looking at traffic engineering (TE) to automate path selection procedures and to maintain network loading at an optimal level. In this paper we propose two TE path selection algorithms that consider the application's delay-bandwidth requirements as well as the TE constraints on the network. We compare the proposed algorithms to existing path computation solutions and present results that show that by considering these additional constraints, improvement is achieved in terms of reduction in request blocking probability, reduction in network costs and load distribution. (C) 2002 Elsevier Science B.V. All rights reserved.