SAFE: Scalable Autonomous Fault-tolerant Ethernet

In this paper, we present a new fault-tolerant Ethernet scheme called SAFE (scalable autonomous fault-tolerant ethernet). SAFE scheme is based on software approach which takes place in layer 2 and layer 3 of the OSIRM. The goal of SAFE is to provide scalability, autonomous fault detection and recovery. SAFE divides a network into several subnets and limits the number of nodes in a subnet. Network can be extended by adding additional subnets. All nodes in a subnet automatically detect faults and perform fail-over by sending and receiving Ethernet based heartbeat each other. For inter-subnet fault recovery, SAFE manages master nodes in each subnet. Master nodes communicate each other using IP packets to exchange the subnet status. We also propose a master election algorithm to recover faults of master node automatically. Proposed SAFE performs efficiently for large scale network and provides fast and autonomous fault recovery.