摘要
事件流描述语言是复旦大学研制的并行分布式计算机监测系统的一部分。为实现系统的通用性(即独立于监测目标与监测方法),设计了测量数据的通用逻辑结构;为支持面向问题的源相关性并提高系统的运行效率,设计了事件流描述语言,用以描述测量数据的逻辑结构,再对该描述文件进行编译,产生了一个二进制访问键文件。该访问键文件通过一个统一的函数接口与系统的分析评估工具相连,具有有效支持监测系统的通用性、面向问题、模块化及易扩充等优点。
The event trace description language is one part of Distributed Monitoring Systems developed by Fudan university. In order to fulfill the requirement of universal(i.e., object system independence and monitor independence), we developed the general logical event trace structure, and to fulfill the requirencmt of problem-oriented source reference, we developed the event trace description language to describe the logical structure of the monitored data. And for efficiency, we compile the description file to access key file which contains a complete description of formats and properties of measured data. The access key file connects the analysis and evaluation tools of the system with a uniform function interface.The system is distinguished by the features of universal, problem-oriented, source reference and extendible.
出处
《计算机工程》
CAS
CSCD
北大核心
1995年第S1期340-344,共5页
Computer Engineering
关键词
并行及分布式监测系统
事件驱动技术
面向问题的分析
parallel and distributed monitoring system / event-drived method / problem-oriented analysis