Research Of Data Allocation Strategy In Distributed Database

The aim of solving fragment allocation is optimizing the total cost of distributed database system. This is also the most important problem in the design of data allocation. For solving the problem of data allocation better, the author of this paper synthesized heuristic remove fragment copies allocation algorithm and heuristic add fragment copies allocation algorithm mainly. Meanwhile, local optimization data allocation algorithm and benefit-cost heuristic fragment allocation method are also adopted. Finally the author proposed a fragment allocation strategy based on the accessing characteristic of data fragment.Statistical information is the basic information for solving the problem of data allocation. In this strategy, the statistical information which will be used in the cost formula is choosed through obeying three principles firstly. Secondly, the cost formula which is mainly about the cost of transaction processing is confirmed. Finally, a fragment allocation algorithm based on the accessing characteristic of data fragment is proposed. This strategy treats data fragment differently by the accessing characteristic and this point is incarnated in every step. This strategy can be divided into tow steps. The first step is making a initial allocation and sorting the data fragments by the ratio of update and read only. The second step is adjusting allocation by the previous order, balancing the entire system cost of update access and read only aceess. Then, it almost can optimize the total cost, that is the total cost of transaction processing is nearly minimal.At the end of this paper, this strategy is validated through experiment. The experiment shows that a better result can be gained by using this strategy than using heuristic remove fragment copies allocation algorithm or heuristic add fragment copies allocation algorithm. Agorithm efficiency is improved because there are some steps which can be executed parallelly. It has been proved that it is simple, comprehensive, practical, and more excellent in collectivity.