海量并行处理系统的大数据读写接口优化研究
Research on interface optimization for reading and writing large data of massive parallel processing system
-
摘要: 针对MPS系统中数据文件管理混乱、文件读写速率低的问题,采用道集流方式进行数据流程控制,并使用超时机制、数据库连接池和内存映射等技术对系统进行优化.测试结果表明,优化后的系统数据管理有序、文件读写速度显著提高.Abstract: Aiming at the problems such as disorderly management of data file and low literacy rate in MPS,the system was optimized adopting the way collecting pattern for data flow control,and using the timeout mechanism,database connection pool and memory mapping technology.Test results showed that data managed orderly and file reading and writing speed increased in the optimized system.
-
Key words:
- database file management /
- way collecting /
- timeout mechanism /
- connection pool /
- memory mapping
计量
- PDF下载量: 20
- 文章访问数: 823
- 引证文献数: 0