CODEWITHSUNDEEP
real-time
Phoenix
Phoenix

Reputation: 123

Differences between lateness, tardiness, laxity, and criticality in real-time systems

In context of Real Time Systems, what is the difference between the terms lateness, tardiness, laxity, and criticality?

Upvotes: 2

Views: 15717

Answers (1)

E. Douglas Jensen
E. Douglas Jensen

Reputation: 531

This is a confused 3-part question. (Why do you not know about Google?)

Part 1 Lateness and tardiness: Lateness is completion time minus deadline; positive lateness is tardiness; negative lateness is earliness.

Part 2: Laxity is also known as slack time. It is the amount of time left after a task completes if the task was started now.

Part 3: Criticality has various definitions depending on the context. In the real-time computing context (Google doesn't help there), it is most commonly used in "time-criticality," safety-criticalty," etc. Informally, it refers to the dependence of a system's desired properties on timeliness, safety, etc. I use the term "time-criticality" as a generalization of "real-time," pertaining to behaviors (e.g., tasks) whose positive or negative utility to the system depends on the timeliness of the behavior. Timeliness is not commonly defined in the real-time computing community but I define it to consist of

the behavior's completion time with respect to the behaviors' scheduling optimality (or more generally satisfiability) criteria (meeting a deadline is a simple special case)

the predictability of the satisfaction (e.g., completion time)--informally, the degree to which the satisfaction can be reasoned about (ideally known) in advance.

Conventional real-time computing focuses almost exclusively on a narrow special case in which satisfiability is simply whether or not all behaviors (tasks) meet their deadlines. (Note that in the very large field of scheduling theory, unfamiliar to the real-time computing community, lateness hence tardiness and earliness are very commonly used in the behavior scheduling satisfiability criteria.) Thus for conventional real-time computing, predictability of satisfiability is the simple special case of knowing absolutely in advance whether or not the deadline-constrained behaviors will all meet their deadlines (give that some very strong presumptions hold).

Upvotes: 7

Related Questions