In this paper, we present a method for utilizing the spare capacity in super-scalar and very long instruction word (VLIW) processors to tolerate functional unit failures. Unlike previous work that was primarily interested in detection of transient faults, we are concerned with more permanent and/or intermittent faults which necessitate processor reconfiguration. Our method utilizes the VLIW compiler or the superscalar scheduler to insert redundant operations whenever idle functional units exist. The results of these redundant operations are used to detect and diagnose functional unit failures. For super-scalar processors, the scheduler can then utilize this information to ensure that operations are performed only on non-faulty units. In VLIW processors, this is equivalent to recompiling the code to run on the remaining non-faulty functional units. Since in certain applications, recompilation may not be possible, we consider two alternative reconfiguration strategies for VLIW processors. These strategies sacrifice storage space and execution time, respectively, in order to reconfigure without recompiling. We present Markov models that describe the behavior of processors using these different approaches and we evaluate their reliabilities. The results show that, while super-scalar and VLIW with recompilation provide the highest reliability, all proposed strategies significantly increase reliability over that of an unprotected processor.

This paper is concerned with evaluating the performability of degradable systems consisting of n primary modules and m spare modules, where all modules are identical. The existence of global spares violates the independence requirement of hierarchical composition modeling which is a common method for dealing with the state explosion problem. We consider the problem of calculating the probability that such a system is in a given performance level at time t when modules have arbitrary failure distribution. All known techniques for solving this problem are suitable only for small numbers of modules. Our analysis shows that the global spares dependence can still be handled satisfactorily using hierarchical composition modeling. Polynomial-time algorithms for calculating the probabilities that the system is in the highest performance level, and in any degraded level in the best and worst cases are presented. The algorithm to find the probability of the highest performance level also provides the first efficient solution for the reliability of non-degradable systems with global spares and arbitrary failure distribution.

Performability is an attribute of a system which combines reliability and performance. Recovery procedures in gracefully-degradable multi-computer systems can have significant influence on performability. A methodology is proposed in this paper to incorporate recovery modeling into performability analysis. This allows the derivation of probabilities for each performance level under different recovery schemes. A general model for n-primary, m-spare multi-computer systems is developed. Also, a general model for pair and spare multi-computer systems is formed in the paper. Several examples are given to demonstrate the methodology. These examples show that the methodology provides more accurate information about the performability of the system being investigated. The use of this methodology, when combined with existing performance evaluation techniques, can help system designers to obtain an accurate evaluation of the performability of systems with recovery and to compare the effect of varying recovery procedures.