Enhanced Variable Splitting Ratio Algorithm for Effective Load Balancing in MPLS Networks

Abstract

Fast load balancing is of the highest importance to utilize the resources effectively and improve the performance of the Multi Protocol Label Switching (MPLS) enabled Internet Protocol (IP) networks. Load balancing splits a part of traffic called splitting ratio from the maximum loaded path in to minimum loaded path. One very strong constrain in balancing the load is the tradeoff between the delay to reach the balanced state and the oscillations generated by splitting the load at the balanced state. To avoid oscillations, the splitting ratio must be kept small. However, smaller splitting ratio takes more number of iterations to balance the load. To overcome this constrain in this study, an enhanced variable splitting ratio based algorithm was proposed to reduce the delay without oscillations. The splitting ratio was varied depending upon the difference between the maximum path load and the minimum path load. A numerical method was used to evaluate the performance of the algorithm and it was applied to a test network. The simulation results showed that the number of iterations required to balance the load was decreased significantly with out oscillations.

Examination of Transmission Quality in the IP Multi-Protocol Label Switching Corporate Networks

“…The router sends the packet to the appropriate interface, adding the new label assigned to the given FEC between the LSR in question and the following router on LDP. The first router in MPLS domain is called as ingress router and the last router in MPLS domain is called egress router [12]. The path of the IP packet in the exemplary MPLS domain has been presented in Fig.…”

Multiobjective routing in multiservice MPLS networks with traffic splitting — A network flow approach

“…A traffic splitting mechanism to improve network security (in case of malicious attacks) and resource utilization in MPLS networks is proposed by Avallone et al (2007). A variable splitting ratio based algorithm is proposed by Murugesan et al (2008), where the load is balanced by an iterative method. In every iteration, a part of the load is split from the maximum cost path to the minimum cost path.…”