IEEE/ACM Transactions on Networking (TON)
Cost-effective traffic grooming in WDM rings
IEEE/ACM Transactions on Networking (TON)
An Ultra-fast Shared Path Protection Scheme - Distributed Partial Information Management, Part II
ICNP '02 Proceedings of the 10th IEEE International Conference on Network Protocols
A novel generic graph model for traffic grooming in heterogeneous WDM mesh networks
IEEE/ACM Transactions on Networking (TON)
Mesh-based Survivable Transport Networks: Options and Strategies for Optical, MPLS, SONET and ATM Networking
Traffic grooming in WDM networks
IEEE Communications Magazine
Traffic grooming in an optical WDM mesh network
IEEE Journal on Selected Areas in Communications
Availability analysis of span-restorable mesh networks
IEEE Journal on Selected Areas in Communications
Hi-index | 0.24 |
This paper investigates the problem of survivable traffic grooming (STG) in shared mesh optical networks and proposes different frameworks for improving the survivability of low speed demands against multiple near simultaneous failures. Spare capacity reprovisioning has recently been considered for improving the overall network restorability in the event of dual failures; here, after the recovery form the first failure, some connections in the network may become unprotected and exposed to new failures. Capacity reprovisioning then allocates protection resources to unprotected and vulnerable connections so that the network can withstand a future failure. In this paper, we propose two different reprovisioning schemes (lightpath level reprovisioning, LLR, and connection level reprovisioning, CLR); they differ in the granularity at which protection resources are reprovisioned. Further, each of these schemes is suitable for a different survivable grooming policy. While LLR provides collective reprovisioning of connections at the lightpath level, CLR reprovisions spare bandwidth for lower speed connections instead. We use simulation methods to study the performance of these schemes under two grooming policies (PAL and PAC), and we show that while CLR reprovisions substantially many more connections than LLR (i.e., potentially more management overhead) CLR yields a much better network robustness to simultaneous failures due to its superior flexibility in using network resources.