Seven Insights into Queueing Theory [pdf]

[rgoldste]

To anyone who finds these models interesting and wants to learn a lot more - I’d highly recommend Mor Harchol-Balter’s textbook: https://www.amazon.com/Performance-Modeling-Design-Computer-.... (She is currently a professor at Carnegie Mellon).

(Self-repost from an earlier queuing theory thread).

[buo]

My favorite book on the subject is Telecommunication Networks: Protocols, Modeling and Analysis by Mischa Schwartz. It's one of the rare books that go very deep into the math, but explains it almost step by step, so it is very accessible.

[colechristensen]

Performance Modeling and Design of Computer Systems: Queueing Theory in Action https://www.amazon.com/dp/1107027500

[cfallin]

Yes, absolutely, seconded! I was fortunate enough to take her class (which she taught using this book) and I left the class feeling like I really understood things, having proved everything starting from basic probability. The rigor is admirable.

[hyperpape]

I think you might have a dead link.

[jph]

Adding these insights to this intro to queueing theory for developers: https://github.com/joelparkerhenderson/queueing_theory

[virtuous_signal]

Despite having a math background, I got stuck on page 1, which is supposed to be definitions, where "Utilization" is used without being defined. With some googling, it turns out this is the ratio of a Poisson parameter to an exponential parameter, under the assumption that arrivals are Poisson distributed and service times are exponentially distributed. Basically, utilization = (average rate of arrivals)*(average service time).

[thraxil]

As it's used on the first page, "utilization" is just the fraction of time that the "service center" is in use. So if it's not doing any work at all, it would be 0, fully utilized would be 1. I don't think you need to connect it to Poisson distributions or go into much more depth to use it to understand the gist of the rest of the paper. I assume that since this is an excerpt from a book, it would've been defined elsewhere.

[User23]

I was surprised to see no mention of Little's law[1]. It's a simple and powerful little theorem with practical implications.

[1] https://en.wikipedia.org/wiki/Little%27s_law

[jihadjihad]

I've seen a few interesting things pop up on HN recently about queueing theory--this post was also neat: https://kottke.org/19/01/its-time-for-some-queueing-theory

[opportune]

Is this not basically https://en.wikipedia.org/wiki/Amdahl%27s_law