A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Information and Communications Technologies, School of Engineering and Technology

Thesis (Ph.D.) - Asian Institute of Technology, 2020

As long as the Internet is a communication portal of human and the limitation network of resources are still resolved, Quality of Service (QoS) is an essential measurement to manage data traffic satisfying the categories of demands. Generally, applications and services require different levels of guarantee in demand of service and network resources. Normally, there are more than one QoS metric on each link for instant delay, jitter and loss rate which have correlation to each other. Link correlation, network state and information are dynamically changed when a data traffic is changed. Feasible path selection which satisfies multiple constraints due to different requirements and dynamic changing of the resources is NPcomplete problem. Several feasible paths can be selected as path candidacy with respect to the best path satisfying QoS metric under consideration. Thus, the path satisfying all QoS metrics with the same significant level is somewhat not realistic and too complicated. This research proposes the algorithm to optimally assigning appropriate weight to each QoS metric by using requirements and slackness (remaining resources) to explore the best path satisfying multi-constrained QoS metrics with different link cost correlations. The proposed algorithm is called General Multi-constrained QoS routing using Weighted Metrics (G_MQW), Zhich adoSWV Whe nonlineaU coVW fXncWion and Uela[ed DijkVWUaœV algoUiWhm. The general mathematical closed-form is derived and the Look-ahead DijkVWUaœV algoUiWhm iV relaxed to accommodate the weighted QoS metrics. To enhance probability of feasible path selection, we propose two control variables of the weight and verify their effects in finding the feasible path satisfying multi-constrained QoS. The path selection in this work is based on only the additive QoS metric for unicast communication in network layer. We evaluate the algorithm by simulation using Matlab on Waxman and Barabási network topologies. The evaluation metrics used are Success Ratio (SR), path optimization and computational complexity. The results are compared with MPLMR, TAMCRA, H_MCOP, MPMP, NLR_MCP and Bidirectional Multi-constrained Routing Algorithm. It is shown that our algorithm, G_MQW, produces the highest Success Ratio with less number of hops under the same environments with the other ones, while the computational complexity is comparable.