Light-weight Monitoring of DB2

The desire to offer on-demand, 24/7 services means there is pressure to quickly identify and resolve problems in a database management system (DBMS). Problem determination tools rely on the existence of sufficient monitoring data to support analysis but monitoring introduces overhead and so causes decreased application performance. The proposed research seeks to provide effective light-weight tools for monitoring and analysis to support problem determination in DBMSs. We propose to view monitoring data as a continuous data stream and to apply algorithms and techniques from data stream mining to the data. In the short term the proposed project will develop a prototype of a problem determination tool that can be of benefit for IBM DB2 customers. In the long term, the proposed project will provide valuable insights into the viability of applying data stream techniques to system monitoring.

Faculty Supervisor:

Dr. Patrick Martin

Student:

Jing Huang

Partner:

IBM Canada

Discipline:

Computer science

Sector:

Information and communications technologies

University:

Queen's University