Research Article
MIDTONE: Multicast in Delay Tolerant Networks
@INPROCEEDINGS{10.1109/CHINACOM.2009.5339828, author={Zunnun Narmawala and Sanjay Srivastava}, title={MIDTONE: Multicast in Delay Tolerant Networks}, proceedings={ChinaCom2009-Wireless Communications and Networking Symposium}, publisher={IEEE}, proceedings_a={CHINACOM2009-WCN}, year={2009}, month={11}, keywords={}, doi={10.1109/CHINACOM.2009.5339828} }
- Zunnun Narmawala
Sanjay Srivastava
Year: 2009
MIDTONE: Multicast in Delay Tolerant Networks
CHINACOM2009-WCN
IEEE
DOI: 10.1109/CHINACOM.2009.5339828
Abstract
Delay Tolerant Networks (DTN) are sparse ad hoc networks in which no contemporaneous path exists between any two nodes in the network most of the time. Due to non-availability of end-to-end paths, multicast protocols of traditional networks fail in DTN because they try to find connected multicast graph between source and destination nodes before forwarding data packets. Routing protocols proposed for DTN follow ‘store-carryforward’ paradigm in which two nodes exchange messages with each other only when they come into contact. In the process, ‘Single-copy’ schemes maintain only one copy of the message in the network at any time and the forwarding node waits for the pre-determined next node to transfer the message. ‘Multi-copy’ schemes spread more than one copy of the message opportunistically when nodes come into contact rather than waiting for pre-determined next node. While Multi-copy schemes improve chances of delivery and work well even without any knowledge of the network, communication overhead and buffer occupancy are quite high for these schemes. We propose Multi-copy routing protocol for multicasting in DTN called “Multicast In Delay TOlerant NEtworks (MIDTONE)” which uses ‘Network coding’ to reduce this overhead without compromising performance. Network coding is a mechanism in which nodes encode two or more incoming packets and forward encoded packets instead of forwarding them as it is. We also propose three novel packet purging schemes to drain packets out of the network which takes advantage of features of network coding to increase buffer efficiency. As simulation results suggest, our protocol achieves significantly less delay to deliver all the packets in infinite buffer case and higher delivery ratio in finite buffer case compared to non-network coding based Multi-copy scheme. We also provide empirical relation to estimate optimal generation size for given network and performance parameters.