[rabbitmq-discuss] Scalable Routing

[Ben Hood]

Martin asked:
> There are units missing in your results. Is it microseconds?

On Fri, Nov 7, 2008 at 2:03 AM, Terry Jones <terry at jon.es> wrote:
> "a linear cost curve will follow a flat line when plotted on a logarithmic scale."

On Fri, Nov 7, 2008 at 8:57 AM, Matthias Radestock <matthias at lshift.net> wrote:
> "each iteration of the benchmark is limited by the average exponent of each
> axis, i.e. (x+y) / 2."
>
> Actually, it's the max of the two. That way you always reach the max
> exponent of both axes.

Thanks to Terry, Martin and Matthias for pointing out bugs in the
article. I've updated it now.

Ben