Research On The Collaboration Query Processing For Cloud Data

With the rapid development of the Chinese market, all kinds of enterprise arises at the historic moment, every enterprise has its own unique confidential data and information. At the same time, the user as a consumer, also needs more demand, which leading to the Chinese market competition being more and more intensified. In face of such situation, enterprises need to partner with enterprises which have common interest, through sharing information, enterprises integrate the data information with join query operations. Due to the privacy of enterprise information, enterprises require a trusted third party at this time, when a user query information, enterprises don’t transmit information between each other, all the data are transmitted to the customer after processing in the third party. Due to this kind of data processing model is different from the traditional query and distributed query, so the conventional query optimization and distributed query optimization is not especially suitable for this kind of system framework. therefore, how to design a solution for this model to optimizing the query processing is urgently to be solved.Firstly, in this paper, aiming at the problem that enterprises can’t transfer data between the nodes in the query model, design a query optimization algorithm for this model. In the query processing, define the priority according to the network transmission cost and the quantity of data of enterprise nodes, then based on the greedy strategy, join the high-priority data operation, finally form the optimal query processing plan. At the same time, in order to deal with the complex network environment, put forward a query adaptive strategy. When data fail to transfer due to abnormal network, change the priority of the data, and deploy new query processing plan.Secondly, for the reason that the third party accept and treat with data from companies, which lead to bigger data need to be treat with, the third party can provide scalable cheap distributed computing ability through the cloud computing platform. In order to optimize the efficiency of query processing, a virtual machine allocation policy that distribution of a query plan to the virtual machine is put forward. Through three factors which are the transmission distance, the CPU utilization and memory utilization, define the strategy execution efficiency, and use multi-objective decision-making for the virtual machine allocation policy set. At the same time, in face of abnormal virtual machine, put forward a corresponding virtual machine migration policies, that migrating query plan from overload virtual machine or abnormal virtual machine to other virtual machine allocation strategy in the set.Finally, in face of the query optimization method and the virtual machine allocation policy proposed in this paper, design a large number of experiments to verify the method and policy, through the analysis of experimental results, it can be seen that the proposed query optimization method and the virtual machine allocation strategy can reduce the time of the query processing, improve the efficiency of query processing, as well as understand the overall performance of the optimized system from the theoretical level.