8 February 2005 Achieving fast and stable failure detection in WDM Networks
Author Affiliations +
In dynamic networks, the failure detection time takes a major part of the convergence time, which is an important network performance index. To detect a node or link failure in the network, traditional protocols, like Hello protocol in OSPF or RSVP, exchanges keep-alive messages between neighboring nodes to keep track of the link/node state. But by default settings, it can get a minimum detection time in the measure of dozens of seconds, which can not meet the demands of fast network convergence and failure recovery. When configuring the related parameters to reduce the detection time, there will be notable instability problems. In this paper, we analyzed the problem and designed a new failure detection algorithm to reduce the network overhead of detection signaling. Through our experiment we found it is effective to enhance the stability by implicitly acknowledge other signaling messages as keep-alive messages. We conducted our proposal and the previous approaches on the ASON test-bed. The experimental results show that our algorithm gives better performances than previous schemes in about an order magnitude reduction of both false failure alarms and queuing delay to other messages, especially under light traffic load.
© (2005) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Donghui Gao, Donghui Gao, Zhiyu Zhou, Zhiyu Zhou, Hanyi Zhang, Hanyi Zhang, } "Achieving fast and stable failure detection in WDM Networks", Proc. SPIE 5626, Network Architectures, Management, and Applications II, (8 February 2005); doi: 10.1117/12.575412; https://doi.org/10.1117/12.575412

Back to Top