多机并发系统动态信息的逆向抽取和过滤策略

针对多机并发系统的复杂性,为了辅助用户从多个层次全面地理解并发系统,就需要抽取出能反映系统整体行为特征的动态信息。从进程的视角研究信息的逆向获取,提出了一种抽取多机并发系统动态信息的方法。此方法基于反射和开放编译的植入机制和共享内存机制,来获取所需要的动态信息,在此基础上运用分类策略来组织、收集和过滤信息,最后进行了系统的实验研究。结果表明,逆向抽取的动态信息能够正确、有效地反映多机并发系统设计时进程级别的关系。

Extraction and filtration of dynamic information for multi-host concurrence system

For the complexity of multi-host concurrence system,in order to help users comprehend a concurrence system at all aspects and levels,it is necessary to reversely extract the dynamic information,which can reflect holistic behavioral features of the concurrence software system.The problem of dynamic information extraction in the process view is studied,and an approach to extracting dynamic information from multi-host concurrence system is brought forward.Based on the instrumental mechanism and shared memory,dynamic information can be obtained.Then,the information is organized,gathered and filtrated by the taxonomic category.Finally,a case study is given to verify this approach.The experimental results show that the extracted dynamic information is correct,effective and can reflect the structure of the multi-host concurrence system in detail at the process level.