Anshuman Tripathi

BACK

Multicasting in Delay Tolerant Networks.

Anshuman Tripathi

Prof Arobinda Gupta, Department of Computer Science and Engineering, IIT Kharagpur

GENERAL : In many cases time is not the only governing parameter for network multicasts. In this project we proposed a novel technique for tree computation intree-based multicast protocols, to compromise on the delivery latency by reducing the message duplication in the network. The project was done as a part of Bachelor's Thesis. In this work we propose a node attribute, called importance, which dictates the probabilistic preference given to a node for selection in a path. By suitably changing the importance of a node during path computations, we show that the number of nodes in a multicast path can be reduced. We studied a generic tree computation algorithm and two of its variants.

END RESULT : The proposed algorithm was able to compute thinner and longer trees as opposed to broader and shallower trees, there by reducing the number of nodes. We have conducted simulation studies with SLAW mobility model for the nodes, and concluded that the proposed algorithm can be fruitfull in reducing the space utilization due to a multicast event.

STATUS : The work has been submitted as a paper in the International Journal of Computing and Network Technology (IJCNT).

POSSIBLE IMPROVEMENTS:
  • Simulation on different types of mobility model can give insigthfull data on the senerios the algorithm works best.
  • Of the proposed generic algorithm, only two variant are studied in this work.
RESOURCES: RESULTS: The proposed algorithm was able to to reduce the number of message copies in the network, by as much as 75%. The algorithm also produced thiner and longer trees to reduce the message copies. This working is shown in some of the figures:

Trees

Original Tree Modified Tree Reduction in number of Nodes Comments
6 Note reduction in duplication
42 Bigger multicast group
The figure below shows the gradual modification of the tree as the parameter for the algorithm is varied.