What This Document Is
This document is a technical paper exploring methodologies for assessing and enhancing the dependability of complex computing systems. Specifically, it delves into the application of automated data collection techniques to measure system and software reliability. It originates from research conducted at Digital Equipment Corporation and focuses on the evolution of reliability challenges within changing computing environments – moving away from centralized models towards distributed client/server architectures. The core focus is on leveraging data to proactively improve product design and operational processes.
Why This Document Matters
This material is valuable for graduate students and professionals in fields like environmental risk analysis, software engineering, systems engineering, and quality assurance. Individuals involved in the design, deployment, or maintenance of large-scale computing infrastructure will find the concepts presented particularly relevant. It’s useful for understanding how to move beyond traditional, often subjective, reliability assessments and embrace data-driven approaches. Those seeking to understand the historical context of reliability engineering and the challenges presented by evolving system architectures will also benefit.
Common Limitations or Challenges
This paper presents a specific case study and methodology developed within a particular corporate context. It does not offer a universal, plug-and-play solution for all reliability challenges. The techniques discussed are focused on *measuring* reliability, and while improvements are implied, the document doesn’t detail specific code fixes or implementation strategies. It also doesn’t cover all possible data collection methods, focusing primarily on on-system logging. It's important to note that the examples used are rooted in the computing landscape of the late 1980s and early 1990s, requiring adaptation for modern systems.
What This Document Provides
* An overview of the shifts in computing environments and their impact on system reliability.
* A discussion of traditional methods for assessing system performance and their inherent limitations.
* An exploration of an automated data collection process for gathering on-system logging information.
* Insights into the analysis of collected data to identify failure causes and areas for improvement.
* A historical perspective on the evolution of reliability concerns in VAX systems.
* Key words and terminology related to system and software reliability measurement.