Gregor Koščak (2010) Traffic flow management analysis in computer data networks. EngD thesis.
Today, computer data networks are being used everywhere around us. To increase reliability most of such networks use redundant links, which may not be fully utilized at all times. The reason lies with the routing protocols, which are used to detect network faults, and at all times direct all traffic only across the best path. By doing this only the optimal path is taken advantage of, while others are idle. High end routers have the capability to redirect part of the traffic across arbitrary path, at administrator’s discretion, while maintaining fault tolerance. These mechanisms are referred to as traffic engineering. They allow to define virtual pathways which are bounded by capacity constraints of the network and demand of the traffic. This thesis compares different simple approaches to achieving the same functionality or at least allowing partial traffic redirection with fault tolerance. It describes approaches with metric modification, traffic load sharing, and use of traffic tunnels and compares these to traffic engineering regarding scalability and configuration complexity.
Actions (login required)