Organization of fault- and failure-tolerant computations in completely connected multicomputer systems (Q1881990)

The paper presents a procedure for the organization of fault-failure-tolerant computations of completely connected multicomputer systems (MCS). The procedure takes into consideration two methods for functional diagnostics and controllable degradation of an MCS and an algorithm for the recovery process in faulty digital computers. The MCSs under consideration include a number of digital computers, each with an individual communication channel with users. The paper can be considered as a reference work in the domain of MCSs.