What is calendar time and execution time in terms of reliability testing?

Question

What is calendar time and execution time in terms of reliability testing?

My book is talking about the Musa Okumoto model and says that it uses calendar time and execution time. However googling and looking into the index is not helping. I see no precedence in the chapter that it is being discussed. It says it uses execution time for determining inter-failure rates (which I know what that is) and calendar time to determine if the reliability (I know this as well) they need is achieved.

Answer 1

I don't know about "in terms of reliability testing", but the meanings of those to terms is as follows:

Calendar time - How long whatever you are testing took in actual time. In other words, if you start a stopwatch when you start a program and check it when the program is done, you are measuring calendar time. This is equivalent to the "real" time reported by the Unix time utility.

Execution time - How much CPU time was used. If there are many things running on the machine at the same time, this could obviously be much less than calendar time. Also, if the program is sitting suspended while it waits for IO, the execution time might be low while calendar time might be high.

In general, users care most about calendar time as they are sitting around waiting for the program to be done. On the other hand, execution time is going to show how much of a load the program is putting on the machine.