Real-Time for the Real World ™
Work-in-progress preview of a work-in-progress book:
Introduction to Fundamentals of Timeliness
and Predictability in Dynamically Real-Time Systems [Jensen 201x]
Informally, a system is a “real-time” one if its core properties of timeliness and predictability of timeliness are integral to its logic, not just performance measures. In general, those properties are dynamic due to intrinsic aleatory and epistemic uncertainties about parameters of the system and its application environment; this book refers to “dynamically real-time.” Despite such uncertainties, dynamically real-time systems have mixed application-specific kinds and degrees of criticality—including even the most extreme safety-critical systems (e.g., for warfare)—which can be expressed in terms of application qualities of services. Traditional real-time computing systems are a narrow special case whose parameters and core properties are predominantly static and periodic—their parameters’ values and time evolution are presumed to be known á priori. Those systems have very limited (albeit potentially important) applicability. Many dynamically real-time systems exist, created by application domain experts outside of, and unseen by, the real-time computing field. These experiences have often been held as enterprise proprietary or government classified. However, real-time systems usually suffer from the lack of a coherent foundation for real-time per se. The design, implementation, and application of real-time systems (both static and dynamic) can be extended and strengthened by creating such a foundation based on first principles for those core properties. This book introduces one approach to that. Timeliness is dynamically expressive using my time/utility (née time/value) functions and utility accrual (née value-based) scheduling paradigm [Jensen 77] [Jensen+ 85]. Predictability of dynamic timeliness requires reasoning about it with some formal theory of uncertainty. This book surveys some popular theories, and focuses on applying mathematical theories of evidence (i.e., Dempster-Shafer theory and its subsequent elaborations). The foundation introduced here has been successfully employed in different real-time contexts having wickedly dynamic parameters and core properties, where traditional static real-time perspectives and traditions were inadequate and counter-productive.
N.B. There still are pages from the previous (c. 2008-2012) version of this site which I have not yet updated and integrated (or removed).