Microfiche. Although It Is Recognized That Certain Portions Are Illegible, It Is Being Released in the Interest of Making Available as Much Information as Possible

If computing system performance is degradable then, as recognized in a number of recent studies, system evaluation must deal simultaneously with aspects of both performance and reliability. One approach is the evaluation of a system's "performability" which, relative to a specified performance variable Y, generally requires solution of the probability distribution function of Y. Prior work on per-formability models and solution methods has focused on the case where Y is discrete; in this paper we consider continuous-valued variables of the type usually addressed in performance evaluation (e.g., average throughput rate, average response time, etc.). The models used are similar to those employed in performance modeling (i.e., Markovian queueing models) but are extended so as to account for variations in structure due to faults. In particular, we consider the modeling of a degradable buffer/multiprocessor system whose performance Y is the (normalized) average throughput rate realized during a bounded interval of time. To avoid known difficulties associated with exact transient solutions, we employ an approximate decomposition of the model, permitting certain submodels to be solved in equilibrium. These solutions are then incorporated in a model with fewer transient states and by solving the latter, we obtain a closed-"form solution of the system's performability. In conclusion, some app'.ications ofi this solution are discussed and illustrated, including an example of design optimization.