Current Proceedings on Technology

Yazarlar: Naotaka Onzuka, Naoki Wakamiya, Masayuki Murata

Konular:-

Anahtar Kelimeler:Routing,Biological model,Control overhead,Computational complexity,Robustness.

Özet: The rapid growth of information networks in size, complexity, and dynamics makes traditional and conventional mechanisms unsuccessful and unfeasible. As an example, OSPF requires complete and consistent information about the whole topology. In addition, its scalability is limited for the high computational complexity. Because of them, it is not possible to perform routing adaptive to network conditions by using dynamic link metrics, e.g. delay and utility. In our former research, we adopted a biologically-inspired nonlinear model, called attractor selection, to realize a highly scalable, robust, and adaptive routing mechanism. It imitates adaptive gene expression of bacteria cells, which autonomously and adaptively synthesize nutrients in accordance with the environmental conditions. The proposal, designed for reactive routing in mobile ad-hoc networks, required the computational cost in the order of and exhibited more robust and adaptive behavior than traditional mechanisms. Therefore, in this paper, we take the same approach to realize robust and adaptive routing in a large-scale wired network. However, routing in a wired network must be proactive rather than reactive. As such, each node has to establish and maintain paths to all the other nodes, which results in the considerable amount of control overhead. Furthremore, in facing to a failure, it takes time to recover. To tackle the problems, we propose methods of overhead reduction and fast recovery. We compare our proposal with OSPF from viewpoints of computational complexity and show that our proposal has lower computational complexity and communication overhead than OSPF while achieving the robustness as high as OSPF.

---