Generation And Optimization Of The Aggregation Query Service Over Larg-scale Stream Data

Research on the theory and technology of stream data began more than ten years ago. Aggregation query is one of the common and important query for stream data. Streaming data sets grow continuously and queries must be evaluated on such unbounded data sets. This brings a lot of difficulties for stream data query. For example, it is not practical to cache all the stream data and execute the query on the cached data. Recently, with the development of the new technology and the outbreak of the sensing data, all kinds of new applications proposed higher requirements for aggregation query over data streams, including generation of aggregation query over data stream as a service, high throughput and low latency processing capacity to reduce the service response delay, which brings new challenges to research on aggregation query over data stream.To tackle these challenges, this paper puts forward an effective aggregation query service over data streams that can reduce the burden of tedious development work on the service consumers. The contribution of this paper includes:① In order to support service usage model, this paper proposes adefinition of aggregation query services over data stream, and presents the basic structure of the aggregation query services over data stream. At the same time, the new aggregation query results will be promptly pushed to the service consumer with push technology.② In order to reduce the service response delay, this paper introduces a kind of method based on the mixed calculation model to optimize the aggregation query services over data stream. The method can get the forecast value of data processing delay according to the demand of calculation by data processing delay prediction model, and then select the appropriate calculation model and the corresponding logic processing method according to the predicted value.③ In order to validate the feasibility of the aggregation query services over data stream and the optimization method, some experments have been designed based on the monitor traffic service over license plate data stream. The experimental results show that the aggregation query services over data stream and the optimization method can push the result and effectively reduce the services response. In this paper, the application system is also designed, and the application case is given based on the data stream of the real license plate, which further verifies the practical value and effectiveness of the aggregation query services over data stream and the optimization method.