Re: Road Map : improved scalability - how to proceed ?

["J.Smith" <temp201@hotmail.com>]

"Kurt D. Zeilenga" wrote :

>
> I note that the community has been "improving scalability"
> of OpenLDAP for years.  Nothing new here.  That is, nothing
> to "start".
>
I wasn't implying in any way that there have not been made significant
improvements in the past, nor that there will not be made any in the future.
However, I was trying to imply that maybe it would be a good idea to focus
the effort somewhat more.


>
> However, I think we need to be careful in regards to use of
> measuring tools.  Such tools should only be used to help
> identity possible problems and to help verify that solutions
> to those problems are reasonable.  I don't think we should
> start chasing benchmark results.  Benchmarks results should
> be taken with a large grain of salt.
>
>From an engineering viewpoint, I couldn't agree with you more here. However,
outside of the community there is a whole other world that is effectively
measuring the 'quality' of products based on the performance measured. For
example, do you remember the fiasco where MindCraft benchmarked the SMB
performance of Linux/Samba vs Windows ? When the entire community cried out
that the benchmarks must have been rigged, just because Windows appeared to
perform significantly better than Linux for file serving ? Where the Red Hat
engineers demanded a recount, and to 'oversee' the quality of the benchmark
performed by MindCraft, and fine-tuned both Linux and SAMBA for optimum
performance ? Only to lead to ... almost identical end results - that
Windows did indeed perform better than Linux/Samba ?

If not, maybe the following URL will refresh your memory ;)
http://www.mindcraft.com/openbenchmark.html

The point im trying to make is just that I don't think that anyone would be
helped by letting something like that happen again. Even though we know
perfectly well that these results should be taken with a large grain of
salt, we also have to realise that there are many people out there who
believe in these reports. So running a few of those benchmarks to see how
you are doing (comparatively) would not cause any mayor harm in my honest
opinion.
;)


Just a thought.


Sincerely,

John Smith