[erlang-questions] Benchmarks

[Thomas Lindgren]

----- Original Message ----
> From: Henning Diedrich <hd2010@REDACTED>
>
> Certainly that's the expected defense. Could you spell that out specifically? 
> And, besides the Yaws / Mochi / Apache shoot outs, are there current benchmarks 
> or any metrics out there that show results of which there is some agreement in 
> the community that they do measure something that IS relevant? And I mean that 
> widely, like, on average 1/4 of code vs. C. That'd be a valid point.
> 

"Four-fold Increase in Productivity and Quality
—Industrial-Strength Functional Programming in Telecom-Class Products
FEmSYS 2001 Deployment on distributed architectures"

http://www.erlang.se/publications/Ulf_Wiger.pdf

(In particular section 7.)

For server benchmarking, I'd also recommend a look at Richard Jones, http://www.metabrew.com/article/a-million-user-comet-application-with-mochiweb-part-1

Judging from this list, performance fairly seldom comes up as an issue, so I guess in the aggregate it's mostly considered good enough.

As a testimonial of sorts, we're using an extended Yaws and Erlang to handle substantial simultaneous upload traffic for an online backup service with quite nice results so far, performancewise. I won't go into details, but I think our current setup will be limited by the available network or disk bandwidth rather than CPU, memory, or any software bottlenecks. So switching to nginx or something like that would just be painful without providing much improved performance.

(And as a bonus, because the servers just chug along ops love it :-)

Best,
Thomas