Skip to Content.
Sympa Menu

grouper-users - Re: [grouper-users] ldap source vs jdbc source

Subject: Grouper Users - Open Discussion List

List archive

Re: [grouper-users] ldap source vs jdbc source


Chronological Thread 
  • From: Tom Zeller <>
  • To: Lynn Garrison <>
  • Cc: "" <>
  • Subject: Re: [grouper-users] ldap source vs jdbc source
  • Date: Thu, 24 Mar 2011 09:03:51 -0500
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:from:date :x-google-sender-auth:message-id:subject:to:cc:content-type; b=E/SS8T7gzshdS1Gf0VD3Mt6exZjhr9h7mtqyu/VHVN37p+NSVTi2rLhCgeWZXZuiuE 0wK89SHZ+Uq5hvq5BCy3zrc7xHSr3mHdYNDR7QNhmgGKJ9GE4TZFEP36bI2gPxhfQQW8 sHIWK+suKAaSeaLFM8TaWFSWMvNlapz27qIDE=

> Querying ldap, for subjects (in grouper) and for member identifiers
> (in ldappcng), probably takes the most time.

Looks like I may be wrong.

I assumed that ldap lookups were the bottleneck, and they may be
depending on the environment, but I found that unnecessary retrievals
of attributes for members, when omitted, resulted in a significant
performance improvement (~80% for a group with 10k members).

The MemberDataConnector returns all attributes for a member (including
memberOf), but if those attributes are not part of provisioning, then
those queries can be skipped. I also found that ignoring the Velocity
template engine used to create ldap search filters resulted in a small
gain.

So, additional configuration options for the MemberDataConnector will
help, which I should create a bug for. Once this is done, then
implementing caching in the SPMLDataConnector and tuning caching in
the CachingSubjectResolver should also help.

TomZ



Archive powered by MHonArc 2.6.16.

Top of Page