Data Diversity: An Approach to Software Fault Tolerance

Crucial computer applications require extremely reliable software. For a typical system, current proof techniques and testing methods cannot guarantee the absence of software faults, but careful use of redundancy may allow the system to tolerate them. Existing methods to provide fault tolerance at execution time rely on redundant software written to the same specifications. Such techniques use design diversity to tolerate residual faults. Diversity in the data space can also provide fault tolerance. This is because program faults often cause failure only under certain special case conditions, and for some applications a program may express its input and internal state in a large number of logically equivalent ways. These observations suggest obtaining a related set of points in the data space, executing the same software on these points, and then employing a decision algorithm to determine system output. Such techniques use datu diversity to tolerate residual faults.