Coding techniques for failure recovery in a distributive modular memory organization
Stephen A. Szygenda, Michael J. Flynn · 1971
This paper considers various coding techniques which could be applied to a memory organization to achieve detection and correction of failures. In this discussion, we define a fault as a malfunction of a systems component and a failure as a manifestation of a fault. Notice that a single fault can result in multiple failures. Thus, techniques such as error-detecting and correction codes, when used alone, are limited in that they operate on failures---not faults.