1988 Reports
Performance Evaluation of Global Reading of Entire Databases
Using simulation and probabilistic analysis, we study the performance of an algorithm to read entire databases with locking concurrency control allowing multiple readers or an exclusive writer. The algorithm runs concurrently with the normal transaction processing (on-the-fly) and locks the entities in the database one by one (incremental). The analysis compares different strategies to resolve the conflicts between the global read algorithm and update. Since the algorithm is parallel in nature, its interference with normal transactions is minimized in parallel and distributed databases. A simulation study shows that one variant of the algorithm can read the entire database with very little overhead and interference with the updates.
Subjects
Files
-
CUCS-336-88.pdf application/pdf 586 KB Download File
More About This Work
- Academic Units
- Computer Science
- Publisher
- Department of Computer Science, Columbia University
- Series
- Columbia University Computer Science Technical Reports, CUCS-336-88
- Published Here
- December 9, 2011