الفهرس | Only 14 pages are availabe for public view |
Abstract With the recent advances in computer communication networks, it is necessary to have efficient hardware and software solutions for network problems. In this thesis, we develop a framework for optimization of network operations based on Software Defined Networking(SDN) and Network Functions Virtualization (NFV).The topologies of networks implemented are the Dragonfly topology and Fat Tree topology. The Fat Tree topology is based on a low-radix network. On the other hand, the Dragonfly topology is a representative of high-radix networks. Low-radix networks have low bandwidth, and hence they give high latency. On the other hand, high-radix networks have large bandwidth, and hence they have lower diameter, latency, and cost. An addressing model is presented for Dragonfly topology and simulated by Mininet. In addition, Dragonfly topology is used with Hamming graphs in data center networks to retain optimal global bandwidth with less groups than the maximum allowed.2D Hamming graph is used to build a balanced Dragonfly topology that does not reach the maximum achievable size for a given router and diameter which can be in the order of millions of nodes. Hence, there is no need for extra cost. The leverage resulted from 2-level dragonfly router designs will extend to multi-level Dragonflies to increase the maximum achievable system size with the same router design. |