Analysis of safety related architectures

Fault and Error latency have a great impact on the dependability properties of control systems for critical applications. The replication techniques that are used to build such systems and the degree of replication usually are tailored to the tolerance of one fault (at a time) and result inadequate to cope with latent errors that show up altogether. For this reason, internal error detection mechanisms are coupled with on-line testing activities (diagnostic tests) intended to stress each component of the system so to induce errors and thus to anticipate their detection (reducing latency). Different testing strategies can be adopted on the basis of the element to be tested, the fault to be 'hunted', the characteristics of the system it is applied to. In this work, we start from this simple consideration and will elaborate on architectural organizations to ensure safe and available service.