A framework for dynamic SLA management under heterogeneous traffic conditions in MPLS networks

The delivery of fast, reliable, and differentiated services with a guaranteed performance level for a constantly increasing number of customers is a great challenge of Internet service providers. This can be made by efficient management of network resources, for which service level agreements (SLAs) are used. This paper presents a framework for dynamic management of SLAs in MPLS networks. In our framework, it is the ingress router of the MPLS backbone, which performs admission control as well as the path selection procedure. For path computation, different advanced routing algorithms are used. The behavior and the performance of these routing algorithms affect the number of users that can be served simultaneously, while delivering a proper performance level, which is agreed in SLA. Therefore, SLA fulfillment is impacted by proper selection of the routing algorithm. In our research we have expatriated the performance and behavior of three routing algorithms for MPLS networks. In order to provide a reality-similar, heterogenous traffic situation, a model with three different traffic classes is used. By means of network simulations, we base our investigation on these SLA parameters: average end-to-end delay, average throughput, average packet dropping probability, average link load and number of utilized links.