Mobility Based Distributed Domain Generation Algorithm for MANETs

Mobility Based Distributed Domain Generation Algorithm for MANETs

Title : Mobility Based Distributed Domain Generation Algorithm for MANETs
Authors :
Manousakis, Kyriakos
Baras, John, S.

Ad hoc networks are very important for scenarios where there is not fixed network infrastructure. These scenarios may appear both in the military and the commercial world. Even though there is much advancement in the area of these networks, they do not scale well. This is due to the inability of the existing protocols (e.g., MAC, routing, security) to tolerate the dynamics of these networks when their size becomes large. A remedy is to apply these protocols in a hierarchical manner. The hierarchy generation in these dynamic environments can be advantageous since the network appears to be smaller due to the aggregation and abstraction resulted from the grouping of nodes. The numerous topological changes can be tolerated easier and the applied protocols can perform better. This division greatly reduces overall overhead (e.g., routing overhead with n nodes goes from O(n2) to O(nlogn)) and allows protocols to be tuned to more homogenous conditions). On the other hand, hierarchy has to be generated carefully in order to be beneficial for the network otherwise, it may harm it, due to the imposed maintenance overhead. Towards this objective we have to take into consideration the network environment and design appropriately the hierarchy generation algorithms. In this paper, we present a mobility based domain generation algorithm (DGA), which in a distributed manner attempts to exploit similarities in the mobility patterns of the participating nodes by utilizing a set of mobility metrics. By grouping together nodes with similar mobility characteristics, the mobility based DGA targets the generation of robust to mobility hierarchical structures. The algorithm presents better scalability and robustness characteristics from well known existing distributed domain generation algorithms.