Mass Discrete Data Fast Retrieval System Design And Implementation

According to the feature of mass discrete data, an efficient retrieval method is stated by referring to the philosophy of distributive database, parallel computing, grid computing and multi-threading.Associating with the research and development of national natural heavy concentrate system, considering the feature of natural heavy concentrate database separately stored in every province, which is of mass data, discrete, irregular distribution, and influenced by hardware factors, such as computer configuration and network transfer speed, distributive database and parallel computing philosophy is applied, zone partition and multi-threading queue method is employed to retrieve and process the discrete data within a certain polygonal zone. The following main outcome and conclusions are achieved:(1)Rational database design could reduce the query overhead of server. In terms of physical design, the optimization of database design is realized by optimizing I/O operation and indexing design, which could improve the query speed remarkably.(2)Optimize the execution performance of SQL statements by utilizing the tools like Query Optimizer of MS SQL Server 2000. Other skills for SQL statements execution optimization are also touched upon.(3)The adoption of partition technique could reduce user waiting time and improve response speed. Partition falls into two categories: database partition and query partition. Database partition could optimize query overhead physically. Query partition, by utilizing parallel computing and grid computing philosophy, could disperse the query operations to reduce system response time.(4)Multi-threading technology could help to realize the CPU load balance between request party and response party. One request party corresponds to multiple response parties. Query requests can be sent to multiple servers by multi-threading, and the messages of request party and response party could be kept consistent by thread synchronization mechanism.(5)Hardware factors, such as network bandwidth, influence the query speed as well. According to the existing network status, utilize some software tools, and update...