The Study And Implementation Of Temporal Data Management Technology Based On Temporal Middleware

With dynamic changes of the real world, information has time-varying properties correspondingly. However, there are many difficulties for the traditional Relational Database (RDB) to express temporal semantics.Temporal database adds two time dimensions on the basis of RDB for storage and management the changes of data state. At the same time, a time variable called "Now" is introduced in order to describe the indefinite deadline of validtime dimension more conveniently. However, its complicated semantics make it difficult to determinate its value in different context and bring the uncertainty of query results as well.At the current time, practices and applications of temporal database relatively hang behind. A considerable direction is to add a temporal data management layer by extending the mainstream relational data model for processing temporal information in application systems. The Middleware named TempDB is a typical representative with the best performance among existing temporal products, but it still provides little support for the valid time indeterminacy.In this paper, a temporal data management demo platform is constructed by using middleware technology, which can deal with the uncertain temporal information. It is the result of combining with excellent thoughts and achievements of scholars and provides ideas and methods in both follow-up expanded study and temporal applications. The main work includes the following two parts:(1) The construction of temporal data management platform based on middleware. In view of data temporality and limitations of traditional RDBMS in handling temporal data, we deeply insight into extensions to non-temporal RDBMS and develop a temporal data management demo platform by integrating algorithms of the temporal middleware called"TempDB", the underlying RDB of which is the lightweight MySQL. On one hand, the proposed platform is compatible with the mature technology of RDB and can support the translation from temporal query language ATSQL2 to standard SQL; On the other hand, it encapsulates details of dealing with time dimensions, making the expression of time more closer to people's thinking, and presenting the feasibility, high efficiency and convenience of Middleware technology in handling temporal information.(2) The study of indeterminate valid time caused by complicated semantics of temporal variable"Now". First, we analyze the multiple semantics of"Now"such as past\present\future semantic and the problems about uncertain temporal query results. And then, we propose an indeterminate temporal data model based on the valid time by introducing several reference time parameters and combining with probability theory aiming at having a qualitative and quantitative judgement on indeterminacy. Finally, we discuss the model's implementation and enhance our platform's support for processing indeterminate valid time.(3) A scenario in personnel management is designed to validate the proposed platform, experimental results show that it is compatible with the standard SQL language and non-temporal business logics, provides support for kinds of temporal operations as well. Hence, it can be widely applied in organizing and managing temporal data.