[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: How wold you go about writing a new OpenLDAP backend?




On 06/02/2017 02:09 PM, Quanah Gibson-Mount wrote:
--On Friday, June 02, 2017 3:02 PM -0400 Prentice Bisbal <pbisbal@pppl.gov> wrote:

Well, that certainly sounds like the missing link, doesn't it? ;)

What is different about your native backends that makes them so much
faster? Domain-specific design? I haven't been keeping up with OpenLDAP
over the past few years. The versions I've used in the bast used BDB for
the backend. I've been charged with upgrading an old LDAP server, and as
I was looking through the documentation for the latest version, I saw
that you have some new native backend(s), but I didn't have time to read
up on them.

You might find some of this useful:

<https://en.wikipedia.org/wiki/Lightning_Memory-Mapped_Database>
<https://mishikal.wordpress.com/2013/05/16/openldap-a-comparison-of-back-mdb-and-back-hdb-performance/>
<http://wiki.zimbra.com/wiki/OpenLDAP_MDB_vs_HDB_performance>

There's also a ton more information at:

<https://symas.com/lightning-memory-mapped-database/>

Scroll to the bottom and look at the "Project presentations" section and the "Benchmarks" section, in addition to the rest of the useful information. ;)

Thanks!

Prentice