SRDE-application of data similarity to process control

The concept of data similarity was introduced by T.-W. Kuo and A.K. Mok (1996) to capture the semantics of real time applications where the consistency requirement of a transaction can be relaxed modulo a binary relation on the data space. Even though the notion of data similarity is naturally suited to many applications in the process control industry, its adoption is hampered by the lack of an easy interface to COTS (commercial-off-the-shelf) components. We report our ongoing work in incorporating the idea of data similarity into embedded applications, especially in the process control industry. Our approach is to allow designers to specify similarity predicates by means of an SQL-like interface to the client applications. Query processing is performed by a compact firm-real-time database engine called SRDE (Similarity-based Real-time Database Engine). SRDE is object oriented and is designed to manage real-time data efficiently in a distributed environment. By exploiting data similarity, SRDE can be used to minimize data traffic according to the client's specifications. SRDE has been implemented in 100% JAVA that supports JDBC API. We demonstrate the use of SRDE in a typical industrial control application.