Due to the time-varying nature of wireless networks, it is required to find robust optimal methods to control the behavior and performance of such networks; however, this is a challenging task since robustness metrics and QoS-based (Quality of service) constraints in a wireless environment are typically highly non-linear and non-convex. This paper explores the possibility of using graph theoretic metrics to provide robustness in a wireless network at the presence of a set of QoS constraints. In particular, we are interested in robust planning of a wireless network for a given demand matrix while preserving end-to-end delay for input demands below a given threshold set. To this end, we show that the upper bound of end-to-end round trip time between two nodes of a network can be approximated by point-to-point network criticality (or resistance distance) of the network. We construct a convex optimization problem to provide a delay-guaranteed jointly optimal allocation of transmit powers and link flows. We show that the solution provides a robust behavior, i.e. it is insensitive to the environmental changes such as wireless link disruption, this is expected because network criticality is a robustness metric. Our framework can be applied to a wide range of SINR (Signal to Interference plus Noise Ratio) values.

This paper presents a novel shared mesh restoration scheme for survivable multilayered networks with Shared Risk Link Groups (SRLG) considerations. We develop a heuristic restoration scheme called Two-Step Partial Protection Risk Algorithm to tackle three major problems: (1) network survivability (2) optimization of spare bandwidth, and (3) offering differentiated classes of service. The proposed algorithm provides partial SRLG-diversity for `high-risk' SRLGs that fail the user-specified availability criteria. A linear program formulation is also derived to minimize the backup bandwidth capacity under flow and link-diversity constraints. We show that fewer demands are disrupted in a SRLG-protected network compared to link-only protection, and the number of blocked demands as well as the disrupted demands depend on the number and length of SRLGs in the network.

This paper presents two novel shared mesh restoration schemes for survivable transport networks. First, we present a two-step load balancing pool sharing (LBPS) scheme and show how this scheme can be used to balance the loads on the network links while still minimizing the reserved working and backup capacity in the network. Then, we propose an iterative load balancing pool sharing (ILBPS) algorithm, designed to eliminate the trap-topology problem associated with all two-step mesh restoration algorithms (including LBPS) while still achieving the LBPS algorithm's goals. We compare the capacity-usage, load balancing, and computation complexity performances of these algorithms with two representative algorithms. We show that with the proposed schemes, the network traffic can similarly, or more evenly, be distributed among network links than with the other schemes, at lower computation cost.

This paper presents two distributed end-to-end shared restoration algorithms in a multi-domain network environment. While the first proposed algorithm yields a pair of ldquolink-disjointedrdquo paths between any given pair of nodes in the network, the second algorithm computes a pair of ldquodomain-disjointedrdquo paths. To address the problem of limited information exchange among the domains, the multi-domain network is topologically aggregated as a single-domain network, called virtual path network, in which each domain is abstracted by its border nodes interconnected by point-to-point virtual paths. The performance of both algorithms are evaluated and compared through simulation experiments.

We present a multi-constrained routing algorithm, called delay-constrained pool sharing (DPS), in a survivable mesh network. The goal of this algorithm is to find a pair of link- disjoint primary and backup paths between a given source and destination nodes, which guarantees full recovery from any single link failure in the network. Our objective is to minimize the resources (such as backup bandwidth) used in the network as well as the total end-to-end delay time along the primary and backup paths. Our approach improves the end-to-end delay performance of the simple pool sharing (SPS) algorithm introduced in the literature, and proposes an integrated approach for quality of service (QoS) and resiliency. Using simulation, we studied the DPS algorithm on the existing North-American transport networks, and showed its improvements compared to the SPS scheme. We also introduce a hybrid algorithm which relaxes the delay constraint on the backup path in favor of maximizing the reusability (sharing) of the backup bandwidth.

We present two novel double-constrained routing algorithms, called: Two-step Delay-constrained Pool Sharing (TDPS) and One-step Delay-constrained Pool Sharing (ODPS), in a survivable mesh network. The goal of both algorithms is to compute a pair of link-disjoint primary and backup paths between a given source and destination nodes, which guarantees full recovery from any single link failure in the network. Our objective is to minimize simultaneously the total end-to-end delay time along the primary and backup paths and the resources (such as backup bandwidth) used in the network. Using simulation, we have studied both TDPS and ODPS algorithms on three existing North-American transport networks. We show that the ODPS scheme eliminates the trap-topology problem associated with many two-step link-disjoint path computation algorithms proposed in the literature. It also outperforms TDPS in terms of the total end-to-end delay along the working and backup paths.

The mainstream dynamic bandwidth allocation (DBA) architectures for Ethernet Passive Optical Networks (EPONs) have employed two independent scheduling mechanisms in order to support quality of service: inter-ONU scheduling (timeslot assignment) and intra-ONU scheduling (priority queuing). These architectures tend to implement the inter-ONU scheduling function at the OLT, whereas the intra-ONU scheduling function at the individual ONUs. Since these scheduling functions have been separated, these architectures cannot generally yield a globally optimized bandwidth allocation. In this paper, a centralized bandwidth allocation model is proposed that implements both the scheduling functions at the OLT. A credit pooling technique is employed that enables the OLT to partition the upstream bandwidth among different class of service queues, and to prevent ONUs from monopolizing the bandwidth. High network utilization is achieved by embedding ONU scheduling decisions in time and by eliminating the channel idle-time overhead, associated with many earlier DBA schemes.

Analytical and simulation techniques are presented to investigate the impact of multiple link failures on the availability performance of class sharing, pool sharing, and dedicated mesh restoration schemes. Formulations are given to synthesize distributed point-to-point traffic that is used in the analysis and simulation studies. It is shown that the availability of backup resources is a function of the number of "troublesome" links along the working paths of demands that share these resources. It is also shown that demands can receive a carrier grade service only when they are routed over very short distances with working paths not exceeding 2 hops.

Metropolitan area and long-haul networks are migrating toward the deployment of optical mesh technologies. This requires, among other things, a new generation of highly intelligent protection and restoration mechanisms to perform functions of protection and bandwidth management. We introduce an architecture that provides differentiated protection services across multiple layers of network hierarchy. A connection at any client layer can request a protection against resource failures at any lower layer. A key aspect of the architecture is the hierarchical tree organization of shared risk link group (SRLG) resources. They represent routing-related failures across all layers of protocol stack. The architecture is very scalable in terms of communicating link-state and bandwidth information between adjacent layers. SRLG trees are used to aggregate this information and provide a summary to the client layer. We discuss the requirements and challenges for routing and signaling mechanisms in order to support the proposed architecture. The complexity of this architecture is evaluated and compared with the complexity of a nonhierarchical alternative.

Through comprehensive simulation studies, we have evaluated the ATM layer end-to-end performance of a constant bit rate (CBR) MPEG2 connection. For most practical cases that we have considered, a cell delay variation (CDV) value of 160 /spl mu/sec is obtained (using 155 Mbps OC-3 rate), well below the 1 msec worst case stated in prior studies. The results also indicate that the uniform distribution proposed in the ATM Forum provides a pessimistic model for the cell delay variation. We demonstrate that a gamma distribution adequately models the cell delay and delay variation (CDV). We also study the effect of various network/source parameters (such as the background load, the number of background sources, the number of hops, and the link-speed) on the variances of CDV and the end-to-end delay of the reference connection. Finally, we examine the variation in the cell delay correlation of a reference connection with the above parameters. We show that the correlation is a decreasing non-periodic/periodic function if the reference CBR is multiplexed with the same/other classes of CBR connections.