Evaluation of error recovery schemes for a high-speed packet switched network: link-by-link versus edge-to-edge schemes

An investigation is made of various error recovery schemes for a very-high-speed packet switched network. The schemes investigated are: (1) edge-to-edge recovery, whereby retransmission of erred packets only takes place between source and destination nodes, (2) link-by-link recovery, whereby retransmission only takes place between adjacent switching nodes, and (3) link-edge-combined recovery, using both link-by-link and edge-to-edge schemes. The performance measures are the distribution of transmission delay and error probability of packets across a network. To obtain these measures, the authors develop a tandem queuing network model with feedbacks, where each queue represents a protocol layer within a switching node rather than a switching node as a whole. Numerical results show that for a network with very-high-speed/low-error-rate channels, the edge-to-edge scheme gives the smallest packet transmission delay, while keeping the packet error probability sufficiently small. The only drawback of the edge-to-edge scheme is that it causes network saturation at smaller network traffic load than the older schemes.<>