Re: openldap libraries & filters

[Jon Roberts <jon@jonanddeb.net>]

Lieuwe Elgersma wrote:
> > The absolute best
> > approach is to perform reads on specific dn's eliminating the need for
> > filters altogether, but of course you need to know the dn's.
> I do know the specific dn's .. but I found that doing say 6000
> separate calls is far more expensive then (in my case as I can't do
> more then 750 at a time) 7 calls with 750 each.

With the 6000 calls, are you doing reads or searches? There is a 
significant difference. Also, are you recreating the connection each 
time? That *is* expensive, even with 7 calls.

I don't know your constraints, so your situation may dictate this 
approach. However, generating the page at the link:

http://www.mentata.com/ds/retrieve/congress/housevote/VC107H21

requires performing over 500 serial reads to a local OpenLDAP 
implementation, and I can assure you any lag you notice is due to the 
internet connection and not the directory performance on the 5 year old, 
sub $1000 server this site currently runs on.

Jon Roberts
www.mentata.com