Event Archive‎ > ‎

An Active Database System Supporting Rule Analysis through Evolving Database States

Active database is a term associated with database systems that provide a mechanism that allows the DBMS to perform certain actions automatically in response to certain events or certain conditions. One of the most difficult aspects of active database technology is the process of monitoring the execution of active databases to support rule analysis and debugging activities. This talk will discuss a solution in the object-oriented paradigm and a case study of the HP SQL/MX implementation of the relational model.

One of the most difficult aspects of active database technology is the process of monitoring the execution of an active database to support rule analysis and debugging activities. Most current active database systems are layered on top of existing database systems, thus limiting the ability to examine the database state and control rule execution. This research presents a built-in architecture for active database systems that uses evolving database states together with an advanced active rule execution model to more readily support the needs of run-time rule analysis and debugging activities. Within this architecture, object deltas are low level primitives that capture changes to individual objects through a system-supported, collapsible type structure that is implemented using object-oriented concepts such as encapsulation and inheritance. Delta abstractions provide an additional layer to the database programmer for organizing object deltas according to different language components that induce database changes. Delta abstractions thus support more effective tools for run-time analysis and debugging of active database rules by providing more meaning to the context in which object deltas have occurred. The rule processing algorithm enables the concurrent execution of active rules using nested transactions. The algorithm is coupled with an execution graph structure that dynamically maintains triggering, priority, and conflict relationships among active rules. Execution graphs allow rule designers to visualize the interactions among active rules as they develop and to influence the flow of execution. Coupled with delta abstractions, execution graphs also provide means for examining the database state during rule execution.

An analysis and debugging tool that utilizes object deltas and their corresponding delta abstractions was built as part of this research in the context of the CDOL (The Comprehensive Declarative Object Language) active, deductive, object-oriented database languages. Results from static analysis tools for termination and confluence properties are used by the tool to focus dynamic analysis activities that are guided by the use of the execution graph. The analysis and debugging tool demonstrates the feasibility of the built-in rule processing architecture as well as the increased capability that it provides for more effective run-time examination of active rule behavior.

About the Speaker

Taoufik Ben Abdellatif

Taoufik graduated from Arizona State University (ASU) with a bachelor in computer systems engineering, master of computer science, and a Ph.D. in computer science. After a brief experience with academia at ASU, he moved to industry, for Compaq, now HP. At Compaq, Taoufik joined the Database Engineering Center group in Austin to work on SQL/MX, a proprietary DBMS running on HP's NonStop Server platform. Taoufik's research interests include heterogeneous and active databases where he has several publications.
Steven Teleki,
Oct 16, 2008, 4:52 PM