Technical Report HW-MACS-TR-0028
|Title||Semantics of Stream Operators for Continuous Queries Over Distributed Data Streams
|Abstract||In recent applications such as financial applications, network management, and sensor monitoring, data that is known as stream data comes in a multiple, rapid and constantly changing manner. Along with the appearance of such stream data, people have a concern about a new kind of query named continuous query, which will constantly be evaluated when the query is alive.
Motivated by increasing demand of information sharing, more requirements come out in data stream field; for example, applying DSMSs (Data Stream Management Systems) into a distributed environment such as Grids. Difficulties arise when use the existing DSMSs that were originally designed for the centralised query processing, e.g., for a continuous query involving aggregation over the sliding windows, the answer will randomly change without any control because there is no clear understanding on the order of the tuples over distributed data streams. In order to handle the problems, we need to reconsider all aspects of data management and processing in the presence of distributed data streams. This consequentially leads to many related research problems in semantics, query language and query execution. The primary contribution of this paper are abstract semantics within both the time series model and the sequence model, and an improved CQL language [AW04, GO03] for defining clear semantics of nested aggregate queries over distributed data streams.|