Fault-injection experiments for distributed objects

摘要

Discusses experiments to study the behavior of distributed objectsin the presence of failures. The work is motivated by a practical needin designing object-based distributed systems. System developers need tounderstand how objects fail and how to handle these failures in theirdesign. We consider two distributed object platforms-DCOM and IONA'sOrbix, an implementation of CORBA. We investigate nine potential failurescenarios. These correspond to three different failure types (hanging,abnormal termination and crashes) of three system components (threads,processes and machines). We design experiments to inject failures intoserver object executions. The results are presented as perceived byclients when these failures occur in the server objects. We apply theresults of these experiments to evaluate the effectiveness of a set ofsimple monitoring and recovery mechanisms and also to suggestimprovements in the current DCOM and Orbix implementations