Issue No. 02 - February (1995 vol. 44)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/12.364537
<p><it>Abstract—</it>This paper addresses the problem of estimating the <it>coverage</it> of a fault tolerance mechanism through statistical processing of observations collected in <it>fault injection</it> experiments. A formal definition of coverage is given in terms of the fault and system activity sets that characterize the input space. Two categories of sampling techniques are considered for coverage estimation: sampling in the whole space and sampling in a space partitioned into <it>classes</it>. The estimators for each technique are compared by means of hypothetical examples. Techniques for <it>early estimations</it> of coverage are then studied. These techniques allow unbiased estimations of coverage to be made before all classes of the sampling space have been tested. Then, the “no-reply” problem that hampers most practical fault-injection experiments is discussed and an <it>a posteriori</it> stratification technique is proposed that allows the scope of incomplete tests to be widened by accounting for available structural information about the target system.</p><p><it>Index Terms—</it>Coverage, fault injection, fault tolerance, estimation, sampling, variance reduction.</p>
Y. Crouzet, E. Martins, J. Arlat and D. Powell, "Estimators for Fault Tolerance Coverage Evaluation," in IEEE Transactions on Computers, vol. 44, no. , pp. 261-274, 1995.