Network resource allocation and provisioning issues: Admission control with contiguous constraint in SONET networks and survivability in packet-over-optical networks.

摘要

In this dissertation, two important issues are addressed, from the network resource allocation and provisioning point of view: admission control in capacitated loss networks with heterogeneous traffic and contiguous resource allocation constraints, and secondly a survivability study in IP-over-Optical networks.; Given a capacitated loss network, with heterogeneous traffic arrivals—in the case that each traffic demand must be assigned in contiguous position, the traffic admission control problem is to find call admission/packing algorithms, such that the objective function will be maximized/minimized. This is an extremely complicated problem in terms of obtaining an optimal solution. Based on the complexity study, it is desirable to find computationally inexpensive heuristic call admission/packing policies to minimize the weighted blocking probability. The First-Fit (FF) and Best-Fit (BF) policy are examined by simulation. The blocking probabilities from these two policies are compared with those of the Complete-Sharing (CS) policy and Optimal-Complete-Partitioning (OCP) policy. A loose optimal lower bound is obtained by applying the theory of Semi-Markov Decision Processes (SMDP). As the first to systematically study this problem, our results (from CS, FF, BF, OCP, SMDP policies) suggest that two novel heuristic admission/packing policies, Best-Fit with Reservation (BFR) and Moving Boundary First-Fit (MBFF), might have higher efficiency.; The second study in this dissertation is devoted to the cost effectiveness of survivability in IP-over-optical networks. A novel joint-layer restoration scheme for IP-over-optical networks is proposed, where the restoration algorithm manages network resources from both the IP layer and the optical layer in a synergistic manner. The essence of this scheme is to attempt to reuse router service ports and the service wavelengths associated with optical switch ports for restoration, rather than reserve additional standby router ports exclusively for network recovery. Mixed integer programming (MIP) optimization problems for this purpose are modeled to determine the restoration paths for affected traffic by any single router failure. Our case studies for service-provider networks show that the joint-layer restoration scheme is much more cost-effective in all cases. The extensive sensitivity analysis concludes that the cost savings are robust.