华中科技大学学报（自然科学版）2011，Vol.39Issue(z1)：15-18,4.
高性能网格工作流中的数据服务优化研究
Optimization of data services in high performance grid workflow
摘要
Abstract
To meet the need of bio-computing, such as high concurrency and large data processing, the data services in CNGrid workflow has to be optimized. According to the experiments, the open-source RDBMS (relational database mangement system) used in original architecture was not well adapted to high concurrent requests and large data processing, and has become the bottleneck of performance. In view of this, we analyzed the actual need of data services in grid workflow, from functionality and performance perspectives, and propose a new data service component; TreapDB, which uses mmap and append only log-data file to reduce the I/O waiting time, and meet the business needs of process data querying. Experiments show that, after the optimization of data services, the overall performance of CNGrid workflow increase greatly, and it has been applied successfully in CNGrid bio-community.关键词
网格；工作流；数据服务；优化；高性能Key words
grid/ workflow/ data service/ optimization/ high performance分类
信息技术与安全科学引用本文复制引用
马强,孙君意,李厚福..高性能网格工作流中的数据服务优化研究[J].华中科技大学学报（自然科学版）,2011,39(z1):15-18,4.基金项目
国家自然科学基金资助项目(90812001). （90812001）
