![]() ![]() ![]() |
![]() |
|
|
![]() ![]() ![]() ![]() ![]() |
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Return to Multi and Federated Database Systems In this work, we present a new approach to control concurrency in multidatabase systems. The proposed approach is based on the use of semantic knowledge to relax the notion of absolute transaction atomicity. Supported by this new concept of atomicity, we propose a new correctness criterion, denoted global semantic serializability, for the execution of concurrent transactions, which provides a high degree of inter-transaction parallelism, ensures consistency of the local databases and preserves autonomy of local databases. Our proposal can also be used to increase concurrency in systems for integrating web data sources based on a mediator mechanism. Two concurrency control protocols we have developed are described. ![]() DiSC'02 © 2003 Association for Computing Machinery |