Loading...

Analysis and Evaluation of Routing in Large-Scale Delay Tolerant Networks

Rashidi, Leila | 2019

531 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 52417 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Movaghar, Ali; Entezari Maleki, Reza
  7. Abstract:
  8. Delay-tolerant networks refer to the networks that lack stable end-to-end paths due to the high mobility of nodes. In order to tackle the intermittent connections, nodes of delay-tolerant networks store the data packets and carry them. Nodes forward data packets according to routing protocols once they enter each other's communication range. In order to increase the probability of delivery, several or many copies of a packet can be disseminated in the network. Over the past years, various routing protocols have been proposed for delay-tolerant networks. These protocols differ from each other in the number of copies that are disseminated in the network or the information used to make routing decisions. Analysis and evaluation of the performance of routing protocols are very important. For example, the results obtained from the performance analysis of routing protocols can be helpful to design efficient protocols, meet the network requirements, and provide the desired quality of service. Since delay tolerant-networks are very dynamic due to the mobility of nodes, analysis and evaluation of the performance of routing protocols are very challenging. The condition becomes worse as the number of nodes increases. For example, if the nodes are heterogeneous, using the Markovian models for performance analysis of networks with a large number of nodes encounters the state space explosion problem. In this dissertation, we first evaluate the performance of the epidemic content retrieval protocol. We propose a monolithic model using stochastic reward net, and then, applying the folded technique, we propose an approximate model in order to tackle the state space explosion problem. This folded model reduces the state space significantly while it has good accuracy to evaluate the average retrieval delay and the average number of transmissions of the epidemic content retrieval scheme. A kind of large-scale delay tolerant networks is non-sparse networks wherein the node density is high. In order to study non-sparse networks, we first analyze the probability of successful delivery until a specific deadline under epidemic routing. We prove that the probability of successful delivery experiences a phase transition as node density increases. Afterwards, we propose three performance models for the epidemic routing in non-sparse delay-tolerant networks. These models are based on ordinary differential equations and a new concept, defined by us, called the pairwise infection rate. The proposed models are validated against the discrete-event simulation. Results obtained from the models and the simulation show that one of the models has good accuracy to evaluate the success probability of the epidemic routing in supercritical networks. Moreover, another model yields good accuracy for both supercritical and subcritical networks
  9. Keywords:
  10. Delay Tolerant Networks (DTN) ; Routing ; Performance Evaluation ; Content Retrieval ; Phase Transition

 Digital Object List

 Bookmark

No TOC