In this paper we discuss approaches to Incident and Problem management within the context of IT Service Management, and its de facto standard the Information Technology Infrastructure Library (ITIL). We show how the Problem Management process attempts to diagnose problem root causes by applying various analysis techniques to historical incident data.
We propose a new categorisation mechanism. We break the data free from its hierarchical categorisation scheme through the use of a free form tagging system. By allowing all system users to categorise incidents using their own terms, we show that while individuals may differ, the aggregate meta data produced for each incident stabilises.
Further, by the application of PageRank analysis to the relationships between tags and incidents, we hope to show useful and interesting correlations. While these may or may not be indicative of a causal relationship, they are nonetheless, new facts about the system under scrutiny.
We conclude by showing the system shows some merit, assuming a certain set of minimum system requirements. If these requirements can be met, then this approach, can become another tool in the system administrators’ arsenal of system analysis approaches.
