Stanford InfoLab Publication Server

Database Support for Efficiently Maintained Derived Data

Adelberg, B. and Kao, B. and Garcia-Molina, H. (1995) Database Support for Efficiently Maintained Derived Data. Technical Report. Stanford InfoLab. (Publication Note: International Conference on Extending Database Technology, EDBT'96, Avignon, France.)




Derived data is maintained in a database system to correlate and summarize base data which record real world facts. As base data changes, derived data needs to be recomputed. A high performance system should execute all these updates and recomputations in a timely fashion so that the data remains fresh and useful, while at the same time executing user transactions quickly. This paper studies the intricate balance between recomputing derived data and transaction execution. Our focus is on effcient recomputation strategies -- how and when recomputations should be done to reduce their cost without jeopardizing data timeliness. We propose the Forced Delay recomputation algorithm and show how it can exploit update locality to improve both data freshness and transaction response time. Keywords: derived data, view maintenance, active database system, transaction scheduling, update locality.

Item Type:Techreport (Technical Report)
Subjects:Computer Science > Active Databases
Related URLs:Project Homepage
ID Code:78
Deposited By:Import Account
Deposited On:25 Feb 2000 16:00
Last Modified:02 Dec 2008 15:34

Download statistics

Repository Staff Only: item control page