Research On The Data Stream Database Management System

In the real world, the data of many applications, such as network control, telecom, electronic business transactions, sensor network etc, are not limited data set that be preserved in storage devices, but data stream. Compared with conventional database whose queries are one-off, data stream database's queries are continuous.The work in this paper can be divided into two parts: theory part and implementation part. In the first part, firstly, a new data stream data model is given. Also, we expand our model in several aspects. Secondly, The use of main memory is an important question in the data stream database system. An approach that determines whether a query is computable or not is given in this paper. An algorithm is presented for determining whether or not any given continuous SPJ query can be evaluated using a bounded amount of memory for all possible instances of the data streams. In the second part, a scheduling algorithm based on combining slack time and value is presented in this paper. The influence of real-time transaction complex relations is sufficiently considered. Furthermore, The algorithm's implementation including transaction acceptance policy and transaction completion/abortion policy is given using three special double-linked lists. The system architecture of GStream is designed according to the current design pattern. In this system, an adaptive concurrent policy is adopted. In the end, we implement part of GStream system in RedHat Linux9.0 OS platform.