Performance analysis of fast distributed link restoration algorithms

Abstract

Four distributed link restoration algorithms are analyzed in detail using a set of important performance metrics and functional characteristics. The functional characteristics are used to explain how these algorithms function and provide insight into their performance. The analysis and simulation results indicate that the Two Prong link restoration algorithm, which is based on issuing aggregate restoration requests from both ends of the disruption and on an intelligent backtracking mechanism, outperforms the other three algorithms in terms of restoration time. The RREACT link restoration algorithm consistently found paths that use fewer spares. I. IntroductionWith the widespread deployment of fiber optic transmission systems and the alarming rate of outages due to fiber cuts [1], there is great interest in strategies for improving the process of restoring disrupted traffic from minutes to sub-seconds following a fiber cut [2]. Automatic protection switching probably is the fastest technique and can switch the disrupted traffic to dedicated spare links in under 50 milliseconds. However, it requires high dedicated spare capacity. With recent advances in digital cross-connect systems, DCS, there is increasing interest in using DCS in network restoration [3,4,5,6,7,8,9,10]. The centralized DCS-based network restoration approach [3,4,5] requires reliable telemetric links between the DCS nodes and the network operation center. It is slower than distributed DCS-based network restoration, where the affected DCS nodes exchange messages directly to restore the disrupted traffic. The hybrid preplanned approach proposed in Bellcore's NETSPAR uses a distributed topology update protocol to identify the fault and then downloads a precomputed routing table according to the fault. The problem with this approach is that the memory required for storing the routing tables is too great [6]. In this paper, we will focus on distributed network restoration algorithms for DCS-based fiber networks.There are two basic approaches to reroute the disrupted traffic due to a fiber span cut. The link restoration approach replaces the affected link segment of a disrupted channel by a spare path between the two disrupted ends. The path restoration approach releases each disrupted channel and 1. This research was supported by MCI with grant #01-80046. 2. John Bicknell was a graduate student at UCCS when this work was done and is now with MCI. lets the source and destination end of the channel re-establish the connection. With the additional release phase the path restoration will take more time than the link restoration. However, the path restoration can find more efficient spare paths with fewer link segments and can handle the node failure situation with the same logic. The network restoration techniques described in [3,5] fall in the path restoration category. In order to achieve fast network restoration, we focus on the link restoration approach in this paper.The existing shortest path algorithms provide the basic building blocks or in...

Capacity Efficiency of Distributed Path Restoration Mechanisms in Optical Mesh Networks

“…Given the advantage of path restoration over link restoration in terms of capacity efficiency, it is surprising to find that relatively little work has been done on path restoration [3,4,5] as compared to link restoration [6,7,8,9,10,11,12,13,14,15]. One promising approach for path restoration is the extension of the Self-Healing Network (SHN) protocol [12] to path restoration introduced by Iraschko and Grover [4].…”