Jun. 5 2007 2:00PM

Jun. 5 2007 3:30PM

The fault-tolerant systems design is aimed to provide the desired functionality even in the presence of difference faults, which is reachable due to use of various techniques of the fault effects recovering. These techniques deal with some properties of self-healing, self-testing, etc., which are concerned, one way or another, with time, which can be spent to discover the faults effect, that is of ``latency" of the faults detection. On the other hand, designers need also in some models to predict the execution time of the tasks for which aimed the target systems. The focus of this talk is the predictive modelling of systems concerning both execution time of correct system and fault latency prediction. The system are modelled as some composition of interacting components, which can be described either as finite state machines (FSM), or as a sequence of some processes execution, which may include parallel-executed activities. Some models of estimation of the latency distribution function based both on random sums theory and Markov chains will be considered.

Pravidelný seminář Inteligentní systémy (Contact person: Tomáš Kroupa)