An On-line Routing Algorithm Based on the Off-line Optimal Computing in MPLS

摘要

Traffic Engineering (TE) is an important mechanism to maximize network resources and provide Quality of Service (QoS) for different Internet applications. Multi Protocol Label Switching (MPLS) enables TE in IP networks. In this paper, we present a new MPLS framework - An on-line routing mechanism based on the off-line optimal computing to guarantee QoS for different types of flow and at the same time, improve the utilization of network resources. The new routing algorithm can select paths for the QoS traffic and best-effort traffic. The framework of traffic engineering appeared in this paper has three components: 1) off-line optimal computing, based on the aggregated traffic matrix, which calculates the pre-allocation of link capacities for the every source-destination pair; 2) getting the path set from the result of the optimal computing, which is performed on-line; 3) on-line routing, which selects paths for the QoS traffic, the QoS traffic will use the shortest path that can meet the bandwidth requirement of the traffic, and best-effort traffic will use the light-load path. The algorithm has been implemented in Network Simulator (NS-2). The simulation results show that the new online routing algorithm significantly outperforms the CSPF (Constrained shortest path first), which greatly increases the throughput of the best-effort traffics, and simultaneously guarantees the bandwidth of the QoS traffics.