Correlated Node Behavior Model based on Semi Markov Process for MANETS
This work addresses network reliability issues in MANETs for applications like military or disaster response, but it is incremental as it extends an existing Node Behavior Model.
The paper tackles the problem of modeling correlated node failures in mobile ad-hoc networks (MANETs) by proposing a Correlated Node Behavior Model based on a semi-Markov process, which quantitatively shows that correlated failures have a much more severe impact on network survivability than other misbehaviors.
This paper introduces a new model for node behavior namely Correlated Node Behavior Model which is an extension of Node Behavior Model. The model adopts semi Markov process in continuous time which clusters the node that has correlation. The key parameter of the process is determined by five probabilistic parameters based on the Markovian model. Computed from the transition probabilities of the semi-Markov process, the node correlation impact on network survivability and resilience can be measure quantitatively. From the result, the quantitative analysis of correlated node behavior on the survivability is obtained through mathematical description, and the effectiveness and rationality of the proposed model are verified through numerical analysis. The analytical results show that the effect from correlated failure nodes on network survivability is much severer than other misbehaviors.