Replica Control Algorithms in Distributed Databases
Abstract
The effects of commutative transactions on distributed database performance was analyzed. The benefits were found to be insignificant unless the number of transactions was large. Two distributed algorithms for adaptive replication of data were developed, one which optimizes the communication cost objective function, and a second which optimizes communication time. A dynamic replication control algorithm was studied and showed improved performance over dynamic noting schemes. Concepts for using replicated data in multilevel secure databases have shown the ability to guarantee one-copy serial ability using a small amount of trusted code. There were 14 articles published under this grant.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1993
- Accession Number
- ADA271571
Entities
People
- Sushil Jajodia
Organizations
- George Mason University