Font Size: a A A

Research On Key Technologies For Data Stream Management

Posted on:2005-09-07Degree:MasterType:Thesis
Country:ChinaCandidate:D F HouFull Text:PDF
GTID:2168360155471865Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
With the development of information processing, the data is modeled as continuous transient data streams rather than as persistent relations in many application areas including financial applications, network traffic monitoring, network security, communication data management, web application, manufacturing, sensor networks. All these are sharing common characteristics of huge volume, unpredictable and sudden arrive, which in turn will challenge the traditional data management methods.Approximation and adaptability are the intrinsic features of data stream process. Since only an approximate but not accurate answer is required in many applications, offering approximate response in time is the main purpose of data stream processing. The adaptability of data stream processing is the key factor influencing approximation.After summarizing a series of problems in the data stream management system in detail, the paper made a further study and discussion on some key techniques including sliding window models, approach of summarize data stream. Moreover the paper studied and summarized the continuous query technique on the base of sliding windows, and analyzed and compared the strategy of processing non-incremental update operator. According to the features in data stream processing, we further lucubrated the functions and characters of sliding window model in data stream processing, based on which, the ideas of hierarchy and multiple time granularities is introduced into the sliding window models, and put forward an approach to modeling hierarchical windows model. The paper probed into the method of non-uniform division on sliding windows, based on which, the multi-resolution data stream summary structure was investigated. On the bases of the division of time domain and value domain, the data stream was described as a two dimension array, and it was illustrated the way of constructing and maintaining multi-resolution data stream summary structure. Finally the method of estimate range query and data distribution was brought out.
Keywords/Search Tags:Data stream management, continuous query technique, sliding windows, hierarchical windows model, multi-resolution data stream summary structure
PDF Full Text Request
Related items