Jaguar: Extending the Predator Database System with JAVA

The Jaguar project is aimed at breaking down the traditional barriers that require SQL query processing to reside on the database server. Indeed, database applications will soon be accessed by large number of clients ranging from Web applications to small-scale personal devices and they will in turn access large collections of data sources ranging from Web servers to mobile sensor devices. In such applications, a large amount of computing resources lie outside the database server they should be utilized for performance and security reasons. The objective of the Jaguar project was to define portable query execution plans that could be executed either on the server, or on a client or on a remote data source a web site, an active disk or a sensor device. Java was chosen as a platform for the execution of these portable execution plans. New techniques supporting the execution of portable query plans on the client-site or on the server-site are the major contributions of the Jaguar project. They have been implemented as extensions to the Cornell Predator object-relational system.