Alex Karasulu a écrit :
Hi folks,
<snip/>
At ApacheCon EU Emmanuel did a presentation with a different setup.
First of all the server could not be tuned properly so it had issues
with memory management and performance would be effected. Secondly
the nature of the experiments were different. The common case where a
bind, simple one entry lookup, followed by an unbind was done.
To be clear, it was : BIND, 10 000 Single entry searches, UNBIND. So the
BIND/UNBIND was issued only once.
The convo we had with Alex yesturday stated the following points :
- We must add some cache tuning configuration on ADS (done)
- I favored out of the box ( not tuning except adding some indexes on
ADS to be equivalent to those used in OpenLDAP) installation of all
products, so I didn't tuned anything.
- My tests reflected - in my mind - the way LDAP servers are used in the
industry : IM or SSO systems use LDAP server to find entries for known
users, so they just do search for a known entry. I may be totally wrong,
of course. Alex consider other scenarios, like Mail user agent, or
yellow pages.
- We must test those two kind of scenarios
To do a fair comparison between products like ADS, FDS and OpenLdap, we
must also define tuning parameters that fit best for each scenarios and
each server, not just do brute force tests out of the box. However, the
ApacheCon EU presentation was just intend to expose ApacheDS overall
performances on a simple scenario and to expose the fact that it's
somehow a usable server, which has some advantages, and, yes, some
drawbacks ;) Not to mention the fact that having done those tests
allowed us to find and fix many problems and points we have to improve.
This was the mean reason we conducted those tests.
Emmanuel
