Correlated survivability analysis model for manets
Mobile ad hoc networks (MANETs) rely on collective nodes effort which requires node to be in cooperative behavior to continuously offer network services. Furthermore, node in MANETs shows correlated node behavior due to topology changes, node misbehavior or security attacks in which poses a signific...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English |
| Published: |
2014
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/14882/1/CORRELATED%20SURVIVABILITY%20ANALYSIS%20MODEL%20FOR%20MANETS%2024pages.pdf http://eprints.utem.edu.my/id/eprint/14882/2/Correlated%20survivability%20analysis%20model%20for%20manets%20.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Mobile ad hoc networks (MANETs) rely on collective nodes effort which requires node to be in cooperative behavior to continuously offer network services. Furthermore, node in MANETs shows correlated node behavior due to topology changes, node misbehavior or security attacks in which poses a significant impact on network survivability. However, correlated node behavior is not reflected as one of the metric in analyzing network survivability with current survivability models. The models did not represent real life scenario with the assumption made on individual node behavior. This limitation resulted inaccuracy when analyzing network survivability. To overcome the limitation of current research, this thesis presents a new network survivability analysis model which captures correlated node behavior to depict node behavior in MANETs and proposed a way to minimize the impact of correlated node behavior. Firstly, before network survivability analysis is modeled, a better understanding of dynamic characteristics of node behavior and its correlated behavior need to be studied and modeled. In this thesis, a merging of semi Markov process and Susceptible-Infection-Remove (SIR) epidemic theory is proposed to stochastically model correlated node behavior. To capture correlated node behavior, correlated degree is proposed in the model as a new metric to measure the impact of network survivability under correlated node behavior. Correlated node behavior model leads to a better understanding and prediction of the critical condition and the speed of spreading correlated node behavior to entire network. Network survivability under correlated node behavior is analyzed based on statistical method of multivariate survival analysis in medical research. The modification of Cox Proportional Hazard regression model in particular correlated hazard function is proposed to analyze the probability of correlated node behavior and to determine variables that significantly influence network survivability. The result on regression analysis shows energy consumption and correlated degree are the most significant variables that influence network survivability. Furthermore, probability of network survivability also can be determined. A new algorithm of topology formation is proposed with correlated degree metric to mitigate the impact of correlated node behavior on network performances. The simulation result shows that, with the new algorithm, energy consumption in MANETs can be balance which prolong node life time and increase network survivability. In addition, new algorithm also prevents network topology from partitioning. With new survivability analysis model, the status of network can be precisely measured and countermeasure can be done earlier to prevent network disruption. |
|---|