The Research And Implementation On Queuing Theory Based On Data Stream Query Processing Model

As the technologies of computer and communication are developing rapidly, the request for information has become more nessary and the information capacity become growing, so the information has become a challenge of information management system. In order to solve the overload problem of information management system, some scholars come up with a solution:before the query processing operations setting a buffer queue to adapt to the outbreak of a stream.Useing queuing theory to establish appropriate buffer queue query model is a solution to the problem of the query queue.Based on a detailed analysis of the existed query queue model, and summing up the advantages and disadvantages of the original query queue model, this paper, sets up a series of new models of the query queue. According to the queuing theory, the probability theory and some other related theories, this paper gives out the derivation process and the expression formula for the average queue waiting time and the average queue length for each model. In addition to the independent query queue model, this article will also put forward themixed operation queue model composed of a number of independent query queue models, and analyze the nature of the model and its influencing factors.In this paper, in order to prove the validity of the queue model and analyze their performance, this paper introduces the deliberate tests run on the theoretical value deduced from each queue model and the analog value from the analog experimental data. In the analog data experiments under the same condition, every group of the analog data is tested by several experiments, so that we can describe the closeness of the two curves more accurately and visually by mapping the CDF function only to prove the accuracy of the theoretical value. About the analog experimental data, this paper analyzes to the important parameters the influence when the strength of the input flows and the service flows. The results show that under certain conditions, the queue models established in this article, as well as the analysis on each queue model are relatively accurate.