Self-tuning Database Operations by Assessing the Importance of Data

Project

Self-tuning Database Operations by Assessing the Importance of Data

Principal Investigator

Boris Glavic

Illinois Institute of Technology

Oracle Fellowship Recipient

Xing Niu, Ziyu Liu

Oracle Principal Investigator

Dieter Gawlick
Kenny Gross, Architect
Paul Sonderegger
Vasudha Krishnaswamy
Zhen Hua Liu

Summary

Data and operational knowledge is of immense value to companies, governments, and end users. However, methods for objectively assessing the value of data have remained elusive. In this project, we develop a framework for objective assessment of the relevance of data and operations and for maintaining such information under evolving workloads. The main enabler of this work are provenance sketches which concisely encode what data is relevant for a query or application. The framework we propose will learn and continuously adapt a model of relevance by intelligently deciding when to collect relevance information for an operation, by clustering operations and data based on correlations, and by learning temporal decay functions for both based on past and predicted future behavior.