PDM中BOM数据的MapReduce遍历查询研究

Research on Traversal of BOMData Based on MapReduce in PDM

物料清单(BOM)是产品数据管理(PDM)系统中最重要的基础数据,层次复杂,版本众多,零部件数量巨大,逐步呈现大数据态.现有基于RDB的BOM查询算法,很难实现高并发的复杂层次查询,且系统运行效率低.通过对BOM数据查询特点的分析,利用云计算技术对BOM查询算法并行化,提出了一种基于MapReduce的BOM数据遍历查询策略.以NoSQL为底层数据存储系统,结合MapReduce并行计算及迭代思想,通过矩阵推演,实现了基于MapReduce模型的BOM查询算法.实验结果表明,所提策略的查询时间随节点域的增加趋于平稳,较传统算法在查询效率上提高了一个数量级,系统性能良好.

The bill of materials （ BOM ） , which has complex hierarchies, numbers of versions, and plenty of parts, is the most important basic data and becomes big data state in the product data management（PDM） environment. The existing BOM query algorithm based on RDB is very difficult to achieve high concurrency of complex hierarchical query requirements and high operating efficiency of the system. In this paper,we analyze the query characteristics of BOM and utilize cloud computing technology to parallel the BOM algo- rithm, and propose a BOM data traversal query strategy based on MapReduce. The strategy uses NoSQL as the underlying storage sys- tem, and combines with the MapReduce parallel computing and iterative methods, through Matrix deduction. The experimental results show that the proposed method is stable with the increase of the node domain and outperforms the traditional algorithm by an order of magnitude, and the system works well.