A generalized model for network survivability

Abstract

The high expectation of network to be available and perform at all time has created growing concerns between network operators and engineers all over the globe. Network should be available to users whenever they want to use them. In the wake of an attack, network should be also able to recover as soon as possible. A survivable network is one that can provide services in the presence of an attack or failure and be able to recover fully in a timely manner. Designing a model for network survivability basically involves the understanding of three key issues, i.e, how the system will function in the wake of failures, what will be the impact of failures on the user, and how to overcome these failures. To achieve these goals, this paper proposes a quantitative approach to evaluate the network survivability in terms of using a generalized model; we consider a more complex system with several faults to study the survivability of a network using Markov chain. In presenting a better survivability performance, some numerical comparisons for the network availability, failure frequency and survivability is also developed.