Fault Representation in Case-Based Reasoning Ha Manh Tran and J¨ urgen Sch¨ onw¨ alder Computer Science, Jacobs University Bremen, Germany {h.tran,j.schoenwaelder} @jacobs-university.de Abstract. Our research aims to assist operators in finding solutions for faults using distributed case-based reasoning. One key operation of the distributed case-based reasoning system is to retrieve similar faults and solutions from various online knowledge sources. In this paper, we pro- pose a multi-vector representation method which employs various seman- tic and feature vectors to exploit the characteristics of faults described in semi-structured data. Experiments show that this method performs well in fault retrieval. Keywords: Case-based Reasoning (CBR), Fault Retrieval, Fault Man- agement, Semantic Search. 1 Introduction Fault management involves detecting, reporting and solving faults in order to keep the communication networks and distributed systems operating effectively. Managing faults in small and homogeneous networks requires not much effort. However, this task becomes a challenge as networks grow in size and heterogene- ity. Proposing solutions for faults not only costs much time and effort but also degrades related network services. Artificial intelligence methods introduce some promising techniques for fault resolution. The Case-based Reasoning (CBR) [1] approach seeks to find solutions for similar problems by exploiting experience. A CBR system draws inferences about a new problem by comparing the problem to similar problems solved previously. The system either classifies a given problem into a group of already known and already solved problems or proposes new solutions by adapting solutions for related problems to the new circumstance of the problem. Existing CBR systems for fault management usually cooperate with trouble ticket systems to take advantage of the trouble ticket database as the case database. These systems only function on the local case database, and thus limit the capability of exploring fault-solving knowledge present at other sites. Using shared knowledge sources, however, not only provides better opportunities to find solutions but also propagates updates in case databases that otherwise frequently become obsolete in environments where software components and offered services change very dynamically. Search engines like Google [2] today furnish information search with global data sources and powerful search techniques. It has become common practice for A. Clemm, L.Z. Granville, and R. Stadler (Eds.): DSOM 2007, LNCS 4785, pp. 50–61, 2007. c  IFIP International Federation for Information Processing 2007