A reliable distributed system using dual level fault tolerance
J.W. Hanna, James D. Johannes · 2003
A reliable distributed environment was developed by beginning with a working distributed environment (POSYBL) that contained no explicitly implemented fault tolerant techniques, and augmenting the system with a dual level of fault tolerance. This created the fault tolerant POSYBL environment. Surveying of several current fault tolerant techniques and estimating the amount of reliability each technique would make towards a reliable distributed system showed that there does not exist a single practical fault tolerant technique that would handle half of a certain set of reliability concerns. Addressing less than half of the set of reliability concerns does not make a system fault tolerant for that set of concerns. Therefore, more than one technique was used to create the fault tolerant POSYBL system, which handles a majority of the listed reliability concerns.>