An Efficient Optimal Reconfiguration Algorithm for FDDI-Based Networks
IEEE Transactions on Parallel and Distributed Systems
Real-time communications in homogeneous and heterogeneous computer networks
Real-time communications in homogeneous and heterogeneous computer networks
An Efficient Fault-Tolerant Multicast Routing Protocol with Core-Based Tree Techniques
IEEE Transactions on Parallel and Distributed Systems
Formalization and correctness of the PALS architectural pattern for distributed real-time systems
Theoretical Computer Science
| Hi-index | 4.10 |
As more organizations field high-speed networks for distributed systems, ensuring the fault tolerance of those networks assumes greater importance. Although the US Department of Defense has adopted the fiber distributed data interface as part of its Safenet standard, FDDI's built-in features do not meet Safenet's reliability requirements. This standard requires that a network survive multiple faults; an FDDI network may fail after just two trunk faults. So how do you manage a real-time, FDDI network to meet such a stringent fault-tolerance requirement? The authors' solution involves the use of online and offline components that manage network resources to work around breaks (faults). An integrated approach, their network management strategy combines temporal and spatial redundancy. Temporal redundancy migrates messages on faulty routes to fault-free ones. Spatial redundancy tries to overcome faults by sending multiple copies of messages, hoping that if one copy gets lost, others will reach their destination. Based on four message characteristics-length, period, deadline, and fault-tolerance limit-algorithms perform network management functions. To minimize network overhead, the authors' management strategy manipulates message sets rather than individual messages. Simulation results show that the integrated approach provides better performance than either a temporal or spatial redundancy approach alone.