Fair Approach Buffer Management Scheme for Delay-Tolerant Network

Abstract

Delay-tolerant networks (DTNs) are mobile networks to cope with extreme environment where may lack of continuous and instantaneous end-to-end network connectivity. This dis-connectivity can be caused by the mobility of the network nods or non-sustained communication capability. The messages are delivered to their respective destinations through store, carry and forward approach in which a node stores incoming messages in buffer, carries them while moving and forwards upon connecting to other nodes. Based on message transmission criterion and the number of message copies, DTN routing protocols can be divided into two categories known as single-copy algorithms (forward based) vulnerable to disruptions and experience unbounded delivery delay. Moreover, when message is lost, then it cannot be recovered. And multiple copy algorithms (flood based) nodes create multiple copies of each message and transmit them to all available connections. As result, message can reach to its destination via multiple paths. Most of the existing buffer management policies, both global and local, focus on either dropping policy or scheduling policy, and most of them compute in an unfair manner, guided by a sole metric parameter blindly and biasedness in a random manner while other important parameter has been totally ignored which means that a fair selection cannot be made. Basing on this point, in this paper a fair approach buffer management policy has been proposed in the situations where there is short contact duration, limited bandwidth and buffer. The proposed policy improves buffer occupancy by implementing the three strategies of buffer management policy namely Scheduling, Dropping and buffer clearing totally for computing purpose using four important local information’s parameters namely Message transmission copy count, time to live, size of messages, time to expired,. The performance of proposed policy is validated through simulation by using opportunistic Network Environment (ONE) simulator. The proposed fair approach buffer management policy (FABMP) & the existing buffer management policies namely default MaxProp buffer management policy (DMPBMP) was applied on MaxProp Routing protocol (MPRP) while Effective & Efficient buffer management policy (EEBMP) applied on Epidemic routing protocol (ERP) for evaluation purpose. They were analyzed on two different metrics namely delivery probability and overhead ratio on two scenarios. The simulation results obtained in this thesis shows that for the proposed policy with (MPRP), the message delivery probability is high and overhead ratio is low in both scenarios. Therefore, the proposed policy of FABMP has better performances than existing policy of DMPBMP and EEBMP by increasing delivery probability and decreasing overhead ratio