Optimal management of dynamic information in Delay Tolerant Networks

  • Authors:
  • Yahui Wu;Su Deng;Hongbin Huang

  • Affiliations:
  • Science and Technology on Information Systems Engineering Laboratory, National University of Defense Technology, Changsha, China 410073;Science and Technology on Information Systems Engineering Laboratory, National University of Defense Technology, Changsha, China 410073;Science and Technology on Information Systems Engineering Laboratory, National University of Defense Technology, Changsha, China 410073

  • Venue:
  • The Journal of Supercomputing
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

To improve the performance of information retrieval in Delay Tolerant Networks (DTN), multiple replicas of information are often deployed. In this paper, we mainly study dynamic information, which evolves with time. For simplicity, we use the information IM as an example. We use the discrete time model, and the time is divided into many slots. In each slot, a new version of the information is created. We say that one node's age is k if it receives a replica of the information IM k驴1 slots ago. Due to the constraint of the buffer or other factors, we cannot deploy too many replicas for a specific message. The goal of this paper is to study efficient ways for distributing the limited replicas to maximize the availability of the information. In particular, we consider two policies, which are source-control policy and destination-control policy, respectively. In the first one, if the source node encounters with one node whose age is k, it will forward a new version of the information to this node with probability p(k). In the second policy, if one node's age is k, it discards the replica with probability p(k). We prove that the optimal value of p(k) in both cases conforms to the threshold form. Simulations based on both synthetic and real motion traces show the accuracy of our model. Numerical results show that the destination-control policy is better.