Connection availability and transient survivability analysis in wireless ad-hoc networks

Abstract

Network survivability performance provides an assessment of how well a network supports its functionalities in adverse conditions. This paper studies a wireless network survivability scenario, in which ad hoc routing failures occur due to node mobility, and it enhances and validates previously published results in the following manner: a) it analyses ad hoc connection availability with the use of a Markovian model, b) it amends a previously published evaluation of mobile ad hoc router "birth" and "death" rates, c) it validates theoretical results with simulations, and d) it uses a transient model to evaluate the impact of connection availability to higher application layers. The overall survivability model combines a steady-state analysis of the Markovian model and a transient analysis of packets queuing for access to the wireless medium. The main assumption for the Markovian model is that changes in the ad hoc routing status follow an exponential distribution with state transitory rates analogous to the number of available routers. These rates also depend on the network geometry and the node mobility model. Given a "random walk" mobility model, it is shown that the connection availability is independent of the node velocities. This is statistically verified by simulations reinforcing the validity of the theoretical Markovian assumption. The impact of link failures to the survivability of the system is studied by emulating the layers between the data traffic source and the wireless MAC protocol with a buffer and, thereafter, by evaluating the amount and duration of transitory buffer overflows. The presented survivability model could serve as groundwork for further research in ad hoc wireless networks, including ad hoc network performance analysis and QoS optimisation.

A Framework to Evaluate the Ability to Provide Essential Services to End Users

“…In summary, models adopted by the existing model-based analysis technol-ogy can be divided into two types: CTMC [12,13,16,17,20,24,25,26] and SM-P [14,15,18,19,21,22]. Both models are just for fixed network topologies, for example, models in [12,16,17,19,20,26] are oriented toward star and flat topologies, models in [13,15,23,24,25] are oriented toward cluster-based structure.…”

“…Both models are just for fixed network topologies, for example, models in [12,16,17,19,20,26] are oriented toward star and flat topologies, models in [13,15,23,24,25] are oriented toward cluster-based structure. However, topology can change anytime such that models no longer apply.…”

“…At present there are two types of evalua-tion approaches of survivability of WSNs: simulation [9][10][11], model-based anal-ysis technology [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26]. The latter usually models the evolution of behavior of WSN as a finite state transition system including Continuous Time Markov Chain(CTMC), Semi-Markov Process(SMP), etc.…”

Systematic Review for Network Survivability Analysis in MANETS

“…et al, 2007). The articles by (Abbagnale, A., Cuomo, F. & Cipollone, E., 2009, Dimitar, T. et al,2004Hekmat, R. & Mieghem, P. Van, 2004;Kalogridis, G., Square, Q. & Lin, R., 2009;Park, S., Song, J.…”