Service Execution Model Generation for Runtime Fault Detection using Message Sniffing
Hwang-Wook Kim, Young-Do Cho, Sooyong Park · Jeongbo gwahaghoe nonmunji. so'peuteuweeo mich eung'yong · 2011
Model-based fault detection using message-sniffing is the most suitable method, which can detect faults of black-box components since but it monitors external interactions among the components not investigates the component's internal. However, it is difficult to construct model for fault detection about complex systems in dynamic environment. Moreover, the constructed model may involve errors and it could not guarantee perfection of the model. To overcome modeling error mentioned above, our approach suggests the method to generate Service Execution Model(SEM) that is constructed using extracted log at runtime by system itself not by developers. For model validation, we apply SEM on service robot named sil-bot 2.0 and analyze the improvement.