分布数据库文件分配的模型及解法

本文根据Chu[Chu 69]提出的FAP课题和分析了十七年以来提出的各种不同的分布数据库文件分配模型和解法,并以Casey提出的模型为兰书,提出了两点假设和四条予分原则,两点假设是:(1)分解后的关系片段之间是逻辑独立的,(2)各节点的容量足够大。四条予分原则是:(1)初始分配原则,(2)初始不分配原则,(3)分配原则,当k_1不为空时,(4)不分配原则,当k_1不为空时。本文提出的模型考虑列文件的分解与并发控制两个因素。并根据实际情况,将动态模型分解为静态模型,从而实现了静态分配。

A Model of File Allocation of a Distributed Database and Its Solution

According to the subject of FAP presented by Chu (Chu 69), this paper analgzes the different models of file allocation of distributed databaes and their solutions presented in recent seventeen years. Based on the model built by Casey, two assumpions and four pre-allocation principles are proposed in model establishment. The two assumptions are. (1) the relational segments after decompositon are logically independent of each other; (2) each node capacity is big enough. The four pre-allocation principles are. (1) initial allocation principle. (2) initial principle of no allocation. (3)allocation principle with an empty Kl, and (4) principle of no allocation with a not empty K1. When the model is presented in this paper, two factors of file decompositio and file concurrent control are taken into consideration. According to practical conditions, the dynamic model is decomposed into static model so as to realize static distribution.