Intermittent Errors Research Articles

Error log analysis: statistical modeling and heuristic trend analysis

Most error-log analysis studies perform a statistical fit to the data assuming a single underlying error process. The authors present the results of an analysis that demonstrates that the log is composed of at least two error processes: transient and intermittent. The mixing of data from multiple processes requires many more events to verify a hypotheses using traditional statistical analysis. Based on the shape of the interarrival time function of the intermittent errors observed from actual error logs, a failure-prediction heuristic, the dispersion frame technique (DFT), is developed. The DFT was implemented in a distributed system for the campus-wide Andrew file system at Carnegie Mellon University. Data collected from 13 file servers over a 22-month period were analyzed using both the DFT and conventional statistical methods. It is shown that the DFT can extract intermittent errors from the error log and uses only one fifth of the error-log entry points required by statistical methods for failure prediction. The DFT achieved a 93.7% success rate in predicting failures in both electromechanical and electronic devices. >

Filtering of discontinuous processes arising in marine integrated navigation systems

A refined stochastic model for the errors of the Loran-C radio navigation aid is described, and it is shown how this model can be used to improve the performance of integrated navigation systems. In addition to the usual propagation errors, Loran-C time of arrival measurements are occasionally plagued with sudden intermittent errors of a particular magnitude and caused by receiver cycle selection errors. These result in sudden large jumps in the calculated position solution. The Loran-C error has been modeled as the sum of a diffusion process, representing the normal propagating errors, and a pure jump process of Poisson type, representing the cycle selection errors. A simple integrated navigation system is then described, based on the Loran-C model and the standard dead reckoning (heading and speed) system model. Assuming that the observed process is governed by a linear stochastic difference equation, a recursive linear unbiased minimum variance filter is developed, from which the Loran-C and dead reckoning errors, and hence position and velocity, can be estimated.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Finite Field Fault-Tolerant Digital Filtering Architectures

Digital filtering architectures that simultaneously offer advantages for VLSI fabrication and contain distributed error control are presented. Such structures require parallelism as well as inherent error- control capabilities because VLSI implementations are susceptible to temporary and intermittent hardware errors. The filtering convolution operation is similar to the formation of cyclicerror-correcting codes so that fault-tolerant systems employing finite field arithmetic may be designed containing such codes imbedded directly in the architecture. The interconnection of such systems produces a fault-tolerant system. In addition, the subsystems possess a common design structure which is easily customized to the particular field required, an attractive feature for yield enhancement. Straightforward realizations depending on parallel algebraic decompositions are studied, introducing the locations for fault tolerance and the role of cyclic codes.

Preface

The transmission links between devices in a computer network may introduce significant errors because the links are analog, whereas computing devices are digital. Moreover, the design error rates for transmission links are often much higher than those tolerated for digital computing devices. Computer hardware and software have been designed to provide recovery when errors occur on communication lines. In the first paper in this issue, Bryant takes this process one step further by continuously monitoring the telephone lines used for data transmission. Line quality monitoring improves network availability by locating line problems before they become critical and by recording the causes of intermittent errors. Experiments at one installation show that use of this method greatly increased circuit availability.

Systems with state re-set

Some computer systems, such as operating systems and communications controllers, have to work continuously for long periods. These systems are almost inevitably affected by intermittent errors, and there is a need to develop design methods which minimise the effect of occasional errors. This paper points out that, if some part of a system can be modelled as a finite state machine, then it is sometimes possible to arrange that this part will reset itself to a correct state after an error. The technique has been used in the design of a simple interactive programming system.

Execution time requirements for encipherment programs

Although encipherment has often been discussed as a means to protect computer data, its costs are not well established. Five experiments were conducted to measure the cpu time on a CDC 6400 required by additive ciphers programmed both in assembly language and in Fortran: a “null transformation” to measure the time to move data without encipherment; encipherment with a one-word key; encipherment with a 125-word key; double key encipherment; and encipherment using a pseudo random key. The results were analyzed for consistency over 100 runs, and the effects of constant and intermittent errors were considered. Timing rates for assembly language encipherment ranged from 498,800 characters per second for a pseudo random key cipher to 2,092,000 characters per second for a constant one-word key cipher. The latter is almost equivalent to the rate required simply to move data without encipherment. Fortran tests required over four times as much cpu time. This paper introduces the idea of enciphering time coefficient the ratio of enciphering time to the time taken to fetch and store data without encipherment.