Computation in faulty stars [hypercube networks]

The question of simulating a completely healthy hypercube with a degraded one (one with some faulty processors) has been considered by several authors. We consider the question for the star-graph interconnection network. With suitable assumptions on the fault probability, there is, with high probability, a bounded distance embedding of K_n×S_n-1 in a degraded S_n , of congestion O(n). By a different method, a congestion O(log(n)) embedding of S, can be obtained. For the hypercube O(1) congestion has been obtained, but this is open for the star graph. Other results presented include a guaranteed O(n) slowdown simulation if there are sufficiently few faults, and upper and lower bounds for the minimal size of a system of faults rendering faulty every m-substar     

Dependability [reliability]

Dependability can be defined by maintenance costs and how long the maintenance takes. It also can be defined in terms of continuous operation. Finally, it can be defined by how well it endures abuse or failure. The following terms help define and quantify dependability: reliability, fault tolerance or performability, maintainability, testability, availability and safety.     

Beating the heat [CMOS hot-carrier reliability]

CMOS hot-carrier reliability at both transistor and circuit levels has been examined. Accurate reliability assessment requires defining suitable criteria for acceptable performance for both circuit and individual transistors. As device designers meet demands for greater speed and more complex circuitry accompanied by shrinking the size of transistor into the deep-submicron regime, they have to contend with increase in current densities and higher electric fields. Though in general a MOSFET's driving capability increases as the channel length decreases, the resulting high field will eventually limit the driving capability of the device. The authors discuss improving CMOS hot-carrier reliability through analysis, modelling and simulation     

Modified `practical Bayes-estimators' [reliability theory]

This paper presents a new formulation of `practical Bayes-estimators' (PBE) for the 2-parameter Weibull model when both parameters are unknown. Overcoming some limitations of the first formulation gave rise to this work, but the results are beyond this intent. These estimators are a tool to improve technical knowledge by using a few experimental data. In this case, the controversy about whether to use Bayes or classical methods is surmounted since estimators, like maximum likelihood, give estimates that often appear unlikely on the basis of technical knowledge of the engineers. A Monte Carlo study supports the following conclusions: if the shape parameter is greater than one, modified PBE maintain the good properties of practical Bayes estimators; otherwise the modified PBE are much better and do not suffer from the past limitation regarding the formulation of the prior interval on the shape parameter itself; and when there are very few data the modified PBE work as a filter that always improves (on average) the prior information if it is poor, or substantially confirms it if it is good. From this viewpoint, Bayes theorem allows statistics to help engineering and not vice versa     

Quality enhances reliability [power supplies]

As deregulation dawns, electric utilities are awakening to the fact that power quality counts with many a customer-and counts for a lot. Even now, the groundwork is being laid worldwide for a new type of service contract between a utility and its large industrial or commercial clients. In brief, the utility promises to supply its customer's operations with power above a certain level of “quality”, and the customer promises to buy power from no-one else for as long as the contract lasts-a valuable concession now that retail wheeling has clients in those categories picking and choosing among suppliers in the electric power industry     

Keeping the lights on [power system reliability]

Of all the energy conversion processes in existence, the US electric power system is the largest and most complex. Unlike industries such as communications and transportation, where a demand in excess of supply produces a “busy signal” or temporary grid lock, the nature of the electric power system is one of instantaneously matching supply and demand. Failure to sustain this balancing act can result in partial or complete breakdown of the grid system. In this paper, the authors describe how maintaining reliable electricity grids in a deregulated power industry will become harder, as temptations to cut corners multiply     

Predicting the reliability of electronic equipment [and prolog]

The use of reliability predictions in the design and operation of electronic equipment has been an evolutionary and very controversial process, and over the past decade, reliability prediction methods have been a focal point for a flurry of books, papers, editorials, opinions, special sessions, and workshops. While it is generally believed that reliability assessment methods should be used to aid in product design and development, the integrity and auditability of the reliability prediction methods have been found to be questionable; in that, the models do not predict field failures, cannot be used for comparative purposes, and present misleading trends and relations. This paper discusses the role of reliability prediction and assessment in design, development, and deployment of electronic equipment; overviews the history of reliability predictions for electronics; discusses the advantages and disadvantages of some current methods; and presents some of the key research questions which need to be addressed     

Calculation of the Poisson cumulative distribution function[reliability applications]

A method for calculating the Poisson CDF (cumulative distribution function) is presented. The method avoids computer underflow and overflow during the process. The computer program uses this technique to calculate the Poisson CDF for arbitrary inputs. An algorithm that determines the Poisson parameter required to yield a specified value of the CDF is presented     

A simple approximation to the renewal function [reliability theory]

The authors present a simple, easy-to-understand approximation to the renewal function that is easy to implement on a personal computer. The key idea is that, for small values of time, the renewal function is almost equal to the cumulative distribution function of the interrenewal time, whereas for larger values of time an asymptotic expansion depending only on the first and second moment of the interrenewal time can be used. The relative error is typically smaller than a few percent for Weibull interrenewal times. The simple approximation methods works very well with one term if not too much accuracy is required (e.g. in the block replacement problem) or if the interrenewal (failure) distribution is not exactly known (e.g. only the first two moments are known). Although the accuracy of the simple approximation can be improved by increasing the number of terms, this strategy is not advocated since speed and simplicity are lost. If high accuracy is required, it is better to use another approximating method (e.g. power series expansion or cubic splines method)     

Shrinkage estimation of threshold parameter of the exponentialdistribution [reliability theory]

The authors study the usual preliminary test estimator of the threshold parameter of the exponential distribution in censored samples. The optimal levels of significance and their corresponding critical values for the preliminary test are obtained. The optimal values of shrinkage coefficients for a preliminary test shrinkage estimator are also obtained on the basis of the minimax regret criterion     

Choosing an optimal set of libraries [software reliability]

This paper presents optimization models for selecting a subset of software libraries, viz, collections of programs, residing on floppy disks or compact disks, available on the market. Each library contains a variety of programs whose reliabilities are assumed to be known. The objective is to maximize the reliability of the computer system subject to a budget constraint on the total cost of the libraries selected. The paper includes six models, each of which applies to a different software structure and assumptions. A detailed branch and bound algorithm for solving one of the six models is described; it contains a simple greedy-procedure for generating an initial solution     

Impact of gate-poly grain structure on the gate-oxide reliability[CMOS]

Time dependent dielectric breakdown of thin oxides, 1.5 to 5.0 nm has been studied for different gate-poly grain structures. The poly grain was varied by the poly deposition, and the source-drain (S/D) rapid thermal anneal (RTA) conditions. The study, which was done on fully fabricated CMOS devices, showed substantial reliability degradation in thin gate oxides (below 2.0 nm), when using S/D RTA temperatures above 1000°C. The results can be explained in terms of the interface roughness at the gate poly interface induced by the S/D RTA temperature above the viscoelastic point of the SiO₂. A possible mechanism for the drastic reliability degradation in thin gate oxides, is the protrusion of poly grains into the softening oxide at high temperature     

A cautionary tale about Weibull analysis [reliability estimation]

This paper makes three points about possible perils of unguarded fitting of Weibull distributions to data: (1) bias is introduced by incomplete data, which may have counter-intuitive effects; (2) bias is introduced into percentile estimates by using regression on log-transformed variables to fit the Weibull parameters, particularly if the percentile to be predicted lies outside the range of the data; and (3) the amount of variation associated with such estimates can be very substantial. A partial solution to the incomplete data problem using simulation is presented, and the maximum likelihood approach to parameter estimation and its advantages relative to regression estimation are explained. The problem arose in predicting life expectancy of long-lived components subject to natural aging which cannot be investigated using accelerated testing and for which the collection of data provides an incomplete life record     

Effect of DOA spares on predicted system reliability [switchingsystem]

The authors examine the effect of the initially defective probability, or as it is also known, the dead-on-arrival (DOA) probability, of spare circuit packs on the predicted reliability of a duplex system. A reliability model to incorporate the DOA probability and the associated spare-pack replacement time is developed. Calculations are made in an attempt to quantify the effect of these two variables on the predicted downtime of a representative duplex system. For the system used in this study, it is shown that if the DOA probability is 5% or more, the predicted downtime estimate is 70% greater than the estimate obtained using the standard assumption of a DOA probability of zero     

Characterizations of Gamma and negative binomial distributions[reliability application]

Characterizations of gamma and negative binomial distributions are presented by using the conditional expectation. A necessary and sufficient condition is given in terms of the failure rate for each distribution. To illustrate the usefulness of the results, the authors discuss the mean residual-life of the gamma distribution     

Performability enhancement of fault-tolerant software

Model-based performability evaluation is used to assess and improve the effectiveness of fault-tolerant software. The evaluation employs a measure that combines quantifications of performance and dependability in a synergistic manner, thus capturing the interaction between these two important attributes. The specific systems evaluated are a basic realization of N-version programming (NVP) (N =3) along with variants thereof. For each system, its corresponding stochastic process model is constructed in two layers, with performance and dependability submodels residing in the lower layer. The evaluation results reveal the extent to which performance, dependability, and performability of a variant are improved relative to the basic NVP system. More generally, the investigation demonstrates that such evaluations are indeed feasible and useful with regard to enhancing software performability     

Estimation of scale parameters for two exponential distributions[reliability theory]

The scale parameter of the exponential distribution is estimated using conditional specification. When there are two censored samples available for estimating the scale parameter, a preliminary test is usually used to determine whether to pool the samples or to use the individual minimum-variance unbiased estimator. This latter estimator (usual preliminary-test estimator) is studied. The optimum levels of significance and their corresponding critical values for the preliminary test are obtained on the basis of the minimax regret criterion. A preliminary-test shrinkage estimator is proposed, and the optimum values of its shrinkage estimator is proposed, and the optimum values of its shrinkage coefficients are obtained. For a mean-square-error criterion of goodness of estimation, the preliminary-test shrinkage estimator is better than the usual preliminary-test estimator     

Performability Evaluation of Fault-Tolerant Satellite Systems

A combined performance and reliability (performability) measure for gracefully degradable fault-tolerant satellite systems is introduced, and a closed-form analytic solution is provided in order to compute the performability for a class of unrepairable systems. An efficient algorithm for the solution of Markovian models which take into account the degradation of more than one type of system components is given. By means of an example on a commercial communication-satellite system we intend to illustrate this novel measure and compare it to more conventional ones.     

An improved minimizing algorithm for sum of disjoint products[reliability theory]

The Abraham-Locks-revised (ALR) sum-of-disjoint products (SDP) algorithm is an efficient method for obtaining a system reliability formula. The author describes a minor modification of the ALR algorithm called the Abraham-Locks-Wilson (ALW) method. The new feature is an alternative method of ordering paths and terms. ALW obtains a shorter disjoint system formula on a test example than any previous SDP method and allows small computational savings in processing large paths of complex networks. As there are different ways to obtain a reliability formula it is useful to use an approach which yields the smallest formula relative to computational effort expended. The extra effort in ordering the terms should be reasonably small and usually leads to improved efficiency in the later stages of the algorithm. ALW allows the analyst to operate in a more efficient way on many problems, particularly if the overlap ordering is used in the early stages of processing but is probably ignored for terms that contain a majority of the Boolean variables