Congestion avoidance in internet protocol (IP) networks

Abstract

As the Internet takes an increasingly central role in our communications infrastructure, the slow convergence of routing protocols after a network failure becomes a growing problem. During failures, the forwarding paths in a network change, the forwarding tables in the routers need to be updated. During the update process, transient link overflows may occur if it is not performed in an orderly manner. Congestion is a problem encountered during transient link updates, therefore to assure congestion avoidance during planned link failures in IP networks, we demonstrate the use of an orderly link cost update algorithm to achieve optimum routing hence avoiding congestion. In this dissertation we present the use of an orderly link cost update method on the Pan European and Germany Network model to avoid congestion during planned link failures. We present a method to obtain the updatable links that do not cause any link overflows. To identify which links need to be orderly updated is a huge permutation problem which ISPs (Internet Service Providers) have been struggling to overcome due to its complexity. Using this method, high load links were established. These are nodes with high link load coupled with high degree order. The objective is to split and move traffic from overloaded links to other links of the network. Using this method, Optimum routing was achieved for the Pan European and Germany network model. The results show that using the proposed link cost update method, there are 6 critical links which can be used to achieve optimum routing. We show that using our approach, we can achieve 100% optimum routing. We use AKITA simulator to implement our algorithm. Keywords: Congestion, Backup Topology, TCP/IP