DSR with link prediction using Pareto distribution
Prashant Singh, Daya Krishan Lobiyal · 2010
In this paper, we have introduced a mathematical model to estimate the availability of the link using Pareto distribution to represent the pdf of epoch length. We have used Pareto distribution since it is assumed that small epoch length will occur more frequently and larger epoch length will occur less frequent. Further, this future estimate of link availability is used in determining alternative routes before the current route fails so that messages are delivered without failures. This timely delivery of data improves QoS by reducing latency, and increasing throughput and delivery ratio by use of alternate route. We have incorporated link prediction model with DSR routing algorithm to test the performance of the model. Further, it has been compared with DSR with link prediction using exponential distribution. The simulation results show DSRP and DSRE improves performance of the network and DSRP performs better than DSRE.