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

RE: back-bdb



> -----Original Message-----
> From: owner-openldap-devel@OpenLDAP.org
> [mailto:owner-openldap-devel@OpenLDAP.org]On Behalf Of Kurt D. Zeilenga

> At 08:55 AM 2001-10-15, Marijn Meijles wrote:
> >> the conversion from database format to in-memory format should
> >> be significantly less than that in back-ldbm.  back-ldbm
> >> uses LDIF in id2entry, back-bdb uses BER.  Optimizing
> >> BER and memory management will shift the balance further.
> >yup. with ldif format it's definitely faster, I did some
> profiling on that.
>
> Numbers would be great....
>
> >Have you already thought about what kind of cache you want to use?
>
> Not yet.
>
> BTW, any help completing back-bdb would be appreciated.  I'm
> swamped with other things at the moment and will likely not
> revisit the code for a few weeks.

I've been using back-bdb as a starting point for a hierarchical backend.
(I copied it to a private "back-hbdb" for lack of imagination.) I've
also got a fast encode/decode method that uses the db buffers in place.
Reading
an entry from the database uses zero additional mallocs; writing an entry
uses
a single malloc.

I also intended to replace DNs with e_id's in the data store; that will add
some overhead to the I/O to map back and forth but it will provide a measure
of referential integrity for not much cost.

I don't know what the ultimate goal is for back-bdb, but I could merge this
in and work on finishing it all up if that sounds like a good idea.
Otherwise
I'll keep it separate.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc