| This dissertation presents an exploration on solving traffic load balancing problems concentrating on the IP layer in hop-by-hop networks. The first part of the traffic load balancing research is focused on bandwidth-sensitive routing for premium class traffic in Differentiated Service (DiffServ) networks, where we assume that the traffic in a network cannot be split. If the traffic in a network is splittable, then the second part of the load balancing research is intra-domain traffic engineering in the network, where the bandwidth-sensitive routing solutions in the first part can be used to achieve better traffic load balancing results.; In hop-by-hop DiffServ networks, the premium class traffic, due to its highest priority, may impose negative inter-class effects onto other classes of traffic. The negative inter-class effects are strongly related to bandwidth utilization on links, leading us to address a bandwidth-sensitive hop-by-hop routing problem called the Optimal Premium Routing (OPR) problem. We prove that the OPR problem is NP-hard. We concentrate on Dijkstra-based heuristic algorithms to provide approximate solutions to the problem. By simulation we show that our Inverse Capacity with Exponential Penalty (ICEP) scheme outperforms other existing heuristic solutions.; Traffic engineering (TE) is crucial in hop-by-hop OSPF networks. We present a novel, edge-based traffic engineering method to approximate the optimal traffic allocation. The new method uses bandwidth-sensitive hop-by-hop routing algorithms as its core. It has three significant advantages. First, it keeps the network core simple and therefore scalable. Second, traffic can be partitioned following flow boundaries at the network edges. Because there is no further splitting in the core, traffic from the same TCP flow is able to travel along the same path, eliminating the out-of-order delivery problem from which other TE approaches may suffer. Finally, since uneven traffic splitting can be implemented at the edges, the new method performs well to approximate the optimal traffic load balancing. |
