Message Drop Control with Priority Buffer Management policy for Delay Tolerant Networks

Abstract

The term delay-tolerant network (DTN) is coined to describe and include all types of discon nected,long delay, disrupted or intermittently-connected networks. So, in such network, commu nication is created with the help of encounter opportunities between nodes and those nodes follow store-carry and forward approach to achieve data delivery. In this approach, each node stores the messages in their buffers until they encounter another node and also to achieve a high delivery ratio most of DTN routing transmits several copies of messages. But the combination of long term storage and several copies of message transmission in this process impose greater transmis sion overhead and buffer occupancy. Most of the previous works on improving the performance of routing do not consider the issue of resource limitations like buffer and bandwidth. However, hand held devices which are having limited storage and bandwidth are used in many of DTN ap plications Therefore, buffer management policies have a direct impact on routing performance of DTNs routing. Basing on this point, in this paper Message Drop control with priority(MDCP) buffer management policy has been proposed in the situations where there is short contact duration, limited bandwidth and buffer. The proposed policy prioritize messages in parallel with controlling unnecessary message drops by using three local information’s namely hop count, time to live and size of messages. The per formance of proposed policy is validated through simulation by using opportunistic Network Envi ronment (ONE) simulator. The proposed Message Drop Control with priority buffer management policy (PMDCP) and existing buffer management policies namely Drop oldest (DO), Drop Largest (DL), Drop Maximum Hop Count (DMaxHop), Efficient buffer management policy (EBMP) was adapted on Epidemic Routing(ER) protocol for evaluation. They were analyzed on two different metrics namely delivery probability and overhead ratio. The simulation results obtained in this thesis shows that for the ER with the proposed policy, the message delivery probability is high and overhead is low. Therefore, the proposed policy has better performances than existing policies by increasing delivery probability and decreasing overhead ratio.