| Distributed database is actually a centralized database and a combination of computer network products, which after a very long period of development. And the characteristics of the distributed database, making it more and more into more use than a centralized focus. This feature is its network features, this feature along with the development of network communication, so that an advantage of distributed databases and more efficient. The quality of the database system's query operation depends on its speed of response time, especially in distributed database systems are not only the speed of response time data processing capability by the computer to determine, and have to calculate the current network speed. Therefore, distributed database systems and general centralized database, cost analysis, not only need to calculate the cost of computer processing and input-output data set with the price, but also computing network communication cost.This paper first introduces the definition of the distributed database system, and a brief description of the structure of the distributed database system. Then introduced the various algorithms to optimize the connection, and on the SDD1 algorithm, CHAIN algorithm and Kruskal made a more detailed introduction. Finally, I study the cognitive multiple query optimization, found that several algorithms are not good use of the geographical characteristics of the distributed database, distributed according to the geographical characteristics of a professional these algorithms can be added parallel computing. So I designed a greedy algorithm based on the query graph, the algorithm is the query graph into multiple query block, then parallel processing of the query block optimization. After the final experimental design, and analysis of experimental data, hoping to prove that the greedy algorithm based on the query graph can reduce the total cost compared to other old algorithm can achieve a breakthrough. |
PDF Full Text Request