Year: 2009
Type: dissertation
Abstract: Most companies relying on an Information Technology (IT) system for their \ndaily operations heavily invest in its maintenance. Tools that monitor network \ntraffic, record anomalies and keep track of the changes that occur in the system \nare usually used. Root cause analysis and change impact analysis are two main \nactivities involved in the management of IT systems. Currently, there exists no \nuniversal model to guide analysts while performing these activities. Although the \nInformation Technology Infrastructure Library (ITIL) provides a guide to the or- \nganization and structure of the tools and processes used to manage IT systems, it \ndoes not provide any models that can be used to implement the required features. \nThis thesis focuses on providing simple and effective models and processes for \nroot cause analysis and change impact analysis through mining useful artifacts \nstored in a Confguration Management Database (CMDB). The CMDB contains \ninformation about the different components in a system, called Confguration Items \n(CIs), as well as the relationships between them. Change reports and incident \nreports are also stored in a CMDB. The result of our work is the Decision support \nfor Root cause Analysis and Change impact Analysis (DRACA) framework which \nsuggests possible root cause(s) of a problem, as well as possible CIs involved in a change set based on di erent proposed models. The contributions of this thesis are \nas follows: \n \n- An exploration of data repositories (CMDBs) that have not been previously \nattempted in the mining software repositories research community. \n \n- A causality model providing decision support for root cause analysis based \non this mined data. \n \n- A process for mining historical change information to suggest CIs for future \nchange sets based on a ranking model. Support and con dence measures are \nused to make the suggestions. \n \n- Empirical results from applying the proposed change impact analysis process \nto industrial data. Our results show that the change sets in the CMDB were \nhighly predictive, and that with a confidence threshold of 80% and a half \nlife of 12 months, an overall recall of 69.8% and a precision of 88.5% were \nachieved. \n \n- An overview of lessons learned from using a CMDB, and the observations we \nmade while working with the CMDB.
Source: UWSpace (University of Waterloo)
Author: Sarah Nadi
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2159500523