Hypercube connected rings: a scalable and fault-tolerant logical topology for optical networks
Computer Communications
| Hi-index | 0.00 |
Several government- and industry-funded prototyping and development activities in WDM optical networks are being conducted at various R&D centers. Optical networks can be deployed on top of an existing physical fiber plant that can provide a transparent all-optical signal path to its upper layer, thereby providing broadband services. For instance, an all-optical layer between the physical and ATM layers in a B-ISDN network will significantly improve both efficiency and flexibility for the provision of broadband services. One attractive feature of optical networks is the ease of its logical reconfigurability, i.e., any desired logical network topology can be embedded on top of any given physical fiber plant, subject to the limitations on the number of available wavelengths and transceivers. The logical topologies can be broadly classified into two categories: arbitrary and regular. Nodal connectivity patterns in regular topologies are very systematic and well defined, which simplifies routing and management operations. However, it is difficult to add an arbitrary number of nodes to a regular topology and still maintain its well-defined structure. Over the past few years, several regular logical topologies have been proposed for optical networks. In this article we provide brief descriptions of these regular logical network topologies and a comparative study of their various performance metrics, e.g., average hop distance (viz. network utilization), routing, fault tolerance, and scalability.