Queuing theory for programmers?

Question

After being burnt several times on things that seemed "obviously fast enough" but sucked performance-wise under load, I'm starting to think that my "gut feeling" might be not enough when doing capacity planning, and some theoretical background is necessary.

So - community, can you point me to good resources on applications of queuing theory to programming?

Whatever - articles, case studies, books.

I found a couple of books that seem to be relevant so far; I'd be happy to hear your opinions on them, if you're familiar:

• Software Performance and Scalability: A Quantitative Approach
• The Art of Computer Systems Performance Analysis: Techniques for Experimental Design, Measurement, Simulation, and Modeling
• Performance by Design: Computer Capacity Planning By Example

Answer 1

"The Art of Computer Systems Performance Analysis: ..." by Raj Jain is excellent and the Bible of software performance testing. (But then I my be biased as I was one of it's (minor) internal reviewers).

However, if you really want to understand this stuff ("The Art..." is more of a survey of methods and principles), then the standards are:

1. Quantitative Systems Performance..., Lazowska, et. al.: The classic standard introduction for Capacity Planning and QNA (Queueing Network Analysis) for whole-systems performance modelling and prediction.

and

2. Performance Solutions: A practical guide..., Smith, etc. Which explains SPE (Software Performance Engineering), which is really what you are trying to do.

I have found both of these to be easy to read and understand, and they are considered the standards of these two practices.

Answer 2

Jain is classic, but there's some good modern stuff too. Assuming you're a mathie, anything by Neil Gunther is good: Guerrilla Capacity Planning is his most recent, preceded by Analyzing Computer System Performance with Perl::PDQ

Bob Sneed and I are working on a more engineering-oriented one, but I'm moving a little slow (;-))

--dave