An Overview of Optical Network Bandwidth and Fault Management

This paper discusses the optical network management issues and identifies potential areas for focused research. A general outline of the main components in optical network management is given and specific problems in GMPLS based model are explained. Later, protection and restoration issues are discussed in the broader context of fault management and the tools developed for fault detection are listed. Optical networks need efficient and reliable protection schemes that restore the communications quickly on the occurrence of faults without causing failure of real-time applications using the network. A holistic approach is required that provides mechanisms for fault detection, rapid restoration and reversion in case of fault resolution. Since the role of SDH/SONET is diminishing, the modern optical networks are poised towards the IP-centric model where high performance IP-MPLS routers manage a core intelligent network of IP over WDM. Fault management schemes are developed for both the IP layer and the WDM layer. Faults can be detected and repaired locally and also through centralized network controller. A hybrid approach works best in detecting the faults where the domain controller verifies the established LSPs in addition to the link tests at the node level. On detecting a fault, rapid restoration can perform localized routing of traffic away from the affected port and link. The traffic may be directed to pre-assigned backup paths that are established as shared or dedicated resources. We examine the protection issues in detail including the choice of layer for protection, implementing protection or restoration, backup path routing, backup resource efficiency, subpath protection, QoS traffic survival and multilayer protection triggers and alarm propagation. The complete protection cycle is described and mechanisms incorporated into RSVP-TE and other protocols for detecting and recording path errors are outlined. In addition, MPLS testbed configuration procedure is outlined with suggested topologies. Open issues in this area are identified and current work is highlighted. It is expected that this paper will serve as a catalyst to accelerate the research and development activities in high speed networking