Maintaining Temporal Consistency Of Embedded Real-time Database

Time is an objective characteristic of the universe, each information should have its corresponding temporal attributes. But in traditional database, the data objects are viewed as"current image", it is considered that the value is an snapshot of the real world entity. However, in some advanced applications, the snapshot is viewed as the current may cause some problems. Data temporality plays an important role in many systems, such as data communication, traffic control, stock exchange and battle command autoimmunization etc.Embedded real-time database system (ARTs-EDB) is an embedded active real-time database management system, which is supported by main-memory database system. So it can ensure no data I/O access when transactions are running. ARTs-EDB was designed to have modularized process structure with multi-thread on the top of an embedded micro-kernel. It can easily be ported to the other embedded operating systems.The time constraint of transaction in ARTs-EDB is reflected by the deadline, it requests that transaction must complete before the deadline, otherwise the effects of transaction are harmful. The time constraint of data object is reflected by the sampling time and temporal validity interval. Until now, the literature on embedded real-time database mainly focused on how to satisfy the deadline of real-time transaction, while that of temporal characteristic of data objects are seldom addressed.Both transaction and data object in ARTs-EDB have obvious time constraints. The correctness of the execution not only depends on the logic results generated by the transaction, but also depends on the generating time. An ARTs-EDB has to maintain consistency of the internal status of a traditional database; the consistency between outside world and its internal status has to be taken into account; temporal consistency of the data object itself and consistency among mutual data objects. Temporal constraints of data object will influence the transaction scheduling and concurrency control.