Re: Update on performance issues with Openldap 2.1.x and Solaris 8

[Adam Williams <awilliam@whitemice.org>]

>>>We are now at 45 queries/second via GSSAPI binds sustained with 18 hosts on 
>>>a single CPU Netra T1.
>>>...
>>>Part of the performance boost came in part with a new 
>>>caching piece for Berkeley DB, so people who are running the 2.1.x series 
>>>with Berkeley DB as their backend, may see an increase in performance if 
>>>they move to 2.1.10 from older versions.
>>Have you compared results to what happens if you run a "real" database?
>>(ideally, comparing both msyql and oracle would be very interesting, I
>>think)
>I've (c.q. "the rest of the world") always maintained that this is
>comparing chalk to cheese. Your two minisculed examples are SQL
>databases. 

Agree.

>Since when has a SQL database been able to propagate parts of
>itself to other SQL databases spread across a network? 

Informix, Oracle, PostgreSQL, and M$-SQL all can do this.  Doesn't make
them more applicable to this situation however.

>Since when has it
>had a hierarchical structure defined by the IETF? 

Yep, take a look at -
http://www.openldap.org/faq/data/cache/378.html

>Since when has it had
>the propensity to do referrals Etc, etc.

This is really a function of slapd itself, not the underlying data
store.

>Who defined SQL as being "a real database?" Why aren't f. ex. BTREE
>databases, ISAM-based databases"real database"s?

True.  I'd wager that bdb would roast Informix/Oracle/etc... for several
kinds of applications anyway.

>>IIRC, openldap supports real database interfaces, doesnt it?

More or less, but it is a bit of a hack.  And mapping a complex data set
really requires views, triggers, and stored procedures in the DB, which
rules out ulta-light engines like sqlite and MySQL.