Date of Award

Spring 5-1-2013

Document Type

Dissertation

Degree Name

Doctor of Science in Information Systems

Department

Business and Information Systems

First Advisor

Surendra Sarnikar

Second Advisor

Amit Deokar

Third Advisor

Stephen Krebsbach

Fourth Advisor

Maureen Murphy

Abstract

Organizations in today’s rapidly evolving digital economy are relying more than ever on their database systems for critical decision-making functions. As a result, speedy and timely availability of the information from these systems is one of key factors crucial to organizational survival. Operating these database systems at high performance levels under highly-integrated, dynamic and complex environments is a knowledge-intensive and an errorprone human-driven task. Although there have been several developments in the area of autonomous performance tuning, such approaches are of limited use because they do not include a holistic view of the problem space and the environment under which they operate. Specifically, these approaches largely ignore the impact and the extent of organizationspecific environmental changes on the performance of their database systems. This research addresses these issues by proposing: 1. A holistic autonomic tuning knowledge model that extends the existing autonomic tuning reference model by incorporating the organizationspecific environmental change impact knowledge. 2. A theory based framework called “DECIPHER” that that not only acquires this knowledge component but does so in a proactive fashion. This framework predicts the potential impact of environmental changes and its dependencies by mining the historical change information stored within the existing organizational incident management data stores. 3. A new change pattern recurrence metric to identify the contexts in which change impact prediction algorithms will be useful and to help identify the best subset of data to use for change impact prediction model building.

Share

COinS