Protected Ethernet Rings for Optical Access Networks

In this paper we propose a centralized link layer architecture for providing low latency fault recovery for optical access rings. This architecture exploits the naturally uneven breakdown of network management responsibilities between the components of an access ring. Important administrative operations like ring status checking, fault detection and recovery are aggregated at the HUB component located in the Central Office of the access network. Consequently, when compared with a standardized Rapid Spanning Tree Protocol implementation, the design of the proposed architecture benefits from a simplified link layer design of the Optical Network Unit, in addition to a significantly reduced fault recovery delay in the ring. We also present an Ethernet-based protocol that realizes our centralized protection model. The design principle of this protocol, responsible for the message passing required to react to topology changes in the network, is simple enough to allow quick reaction times, and to support QoS-aware prioritization of network traffic. The performance of the proposed architecture is evaluated using analytical and simulative means, and the performance aspects of the ring protocol relevant to network protection are compared with those provided by the Rapid Spanning Tree Protocol.