| Summary: | 碩士 === 逢甲大學 === 通訊工程所 === 96 === The major researches in the optical backbone network are the “Fault Diagnosis”,“Fault Location”, and “Fault Recovery”. However, this paper aimed at the “FaultLocalization”, and proposed the corresponding diagnosis method, to promote the efficiency of operation by deep discussion and analysis. Generally, an ISP may have tens of thousands of equipments, including optical fiber, optical amplifier, and optical cross-connect, etc. Someone had already proved this is a NP-complete question [1] if we want to design a perfect multi-failure location algorithm. Supposing we insist to do the system, we will pay the expense for more and more monitoring quipments. Therefore, we aimed to improve the ratio of the diagnosis and the efficiency of the computation so that we can make the balance between the cost and the efficiency.
In consideration of above, we proposed Fault Location Algorithm with SRLG (FLAG). In general, there are more than four cables that dispose on the same channel when fiber construction. Therefore, the risk of channel destruction will be shared by the cables. We named this situation Shared Risk Link Groups (SRLG). For the SRLG information, we stored it in the dependency matrix. We just only diagnose the relationship between the alarm and the dependency matrix, and then we can get the location of failure. Moreover, the dependency matrix can be diagnosed for multifailure through extending and grouping steps. In addition, we proposed the method to sort repair sequence for solving failure concealment with uncompleted information. In the past, the maintenance worker must guessed the failure location depend on experience rules. In our method, we provide the worker the suspected group to reduce repair time. Nevertheless, we adopt pre-computation phase (PCP) for efficiently computing. That is, the part of the computation will be done before the system on line. The remainder of the work will be coordinated to diagnose the failure location by alarming mechanism and diagnosis system. This means that we can diagnose the
failure location in a short time, and do not need expend much for constructing
diagnosis information. In the summarization, the contributions of FLAG are SRLG diagnosis, locating multi-failure, and providing repair sequence. As the above, the essential of this paper has a focus of attention on how to use SRLG information to diagnose multi-failure, and provides a repair sequence table. Finally, we will use the examples to simulate the effect of FLAG, and address the conclusion and future work.
|
|---|
