Robust topology management in DT-MANETs: An efficient tree-based decentralized and multi-objective approach

Doctoral thesis (2010)

Delay-tolerant mobile ad hoc network (DT-MANETs) feature frequent and long duration partitioned MANETs. It is a challenged environment where end-to-end connectivity cannot always be obtained. Furthermore ... [more ▼]

Delay-tolerant mobile ad hoc network (DT-MANETs) feature frequent and long duration partitioned MANETs. It is a challenged environment where end-to-end connectivity cannot always be obtained. Furthermore, communication in such network is heavily relying on collaboration between nodes since there is no central authority. Limited resources of communication nodes present another facet of problems in such network. Moreover, the communication is typically done using wireless technologies which are sharing among communication nodes. In order to provide a better quality of service in such environment, topology management technique is used to help controlling network topology. The aim is to manage network resource and to enhance efficiency of communication. In this work, we proposed to do it by constructing an efficient and robust tree-based topology. We model the environment using dynamic and partitioned graph. Under such circumstances, protocols must withstand topology and condition changing. In summary, doing topology management in DT-MANETs encounters the following issues: cooperation among nodes, limitation of resources of mobile node, sharing medium, dynamic and partitioned topology and unrealistic and unscalable approach of centralized algorithm in such environment. This study focuses on managing tree-based topology in DT-MANETs. A set of active links is managed such that the deterioration of network is avoid and the quality of service in DT-MANETs is enhanced. Efficiency and robustness metrics are proposed accordingly. This work proposed to use different quality criteria based on communication node and edge for the selection topology. Both single- and multi-objectives tree-based topology are studied. Algorithms for managing tree-based topology are proposed according to different objectives. All proposed algorithms are purely decentralized, asynchronous algorithm and use one-hop information. [less ▲]