Simulation of resilient networks

Abstract

Resilient network is a network, which does not fail under any circumstances. It is the ability of the network to provide and maintain an acceptable level of service in the face of various challenges to normal operation. Disjoint routing strategies are used to calculate disjoint paths, which can be used by the network to route the packet in case of some failure to the existing primary path. Packets arrive at a node in the network in a very random manner. The probability density function for describing the number of such arrivals during a specific period follows the Poisson distribution. The inter-arrival time and the service time at a node follow exponential distribution. As a packet travels from one node to the subsequent node along its path, the packet suffers from different types of delays at each node along the path. Thus the total node delay can be calculated by summing up all the different types of delays. With the increase in traffic and therefore the congestion, the average cost of transmission and retransmission (in the event of a failure) suffers. This varies according to the strategy used for resilience. Routing protocols inherently provide a basic level of resiliency. The ability to “route” around the problem areas defines the efficiency of a routing protocol. However, the time to re-converge the network can vary greatly depending on the protocol being used. When a failure occurs, either the packets can be resent or a new path can be followed from the failure point.