An ad hoc network consists of mobile nodes (hosts) that are equipped with wireless transmitters and receivers, which allow them to communicate without the help of wired base stations. With the increasing acceptance of the ad hoc networks the demand for guaranteeing QoS to support various real-time applications has become a major research are. Guaranteeing QoS in ad hoc networks poses a challenging problem, because of the constantly changing network scenario along with the ever changing the QoS state information associated with nodes and bringing in imprecision about the available state information.
In our attempt to propose a solution for QoS routing in ad hoc networks we approach this problem by first proposing an efficient routing algorithm and then incorporate QoS routing into the algorithm. The main thrust in our routing algorithm is to reduce the indeterminism factor by characterizing the node movements to various mobility patterns, which can be expressed by mathematically bound functions. Thus in our attempt to quantify the randomness of the motion of mobile nodes, we try to approximate the movements of the mobile nodes and the network as a whole to one of the defined mobility patterns. Then use this information pre-compute routes to a destination before the current link is snapped as now we can predict the possible future locations of the nodes, we can attempt to guarantee QoS as now can also predict the state of a link. Thus with the knowledge of mobility pattern we are in a position to save lot of overhead which the routing algorithm could have incurred either due to a new route initiation process in the event of route failure, in addition to that now we also don’t have to buffer the packets which otherwise had to be buffered at an intermediate node till the new route was computed or in the process of taking collecting the status of each node in case of QoS routing.
To sum up our algorithm has four key features 1) establishment of low cost path satisfying the required QoS; 2) the route discovery process comes up with multiple paths which are to be used in case of the failure of the primary path and thus increasing the robustness; 3) the path discovered are more robust as the algorithm takes into consideration stationary and transient links and 4) as our algorithm is not a table driven one it is highly scalable.