Re: LDAP/x.500 Search Benefits??:RE

Jochen Keutel (
Tue, 24 Sep 1996 08:58:44 +0200 (MESZ)


> >Again - it depends on the server. I'm currently experiencing with 1 million
> >entries in one X.500 server - certain searches are as fast as the
> >corrosponding SQL query in the original (Informix) database.
> And certain searches are ??
> a) half as fast
> b) 1/10th as fast
> c) so slow I've never waited for them to complete.

I don't know exactly. It's not my job to compare X.500 and SQL. (It would be
an interesting job, of course ...) All I have is the X.500 server -
and I try to compare the results with some measurements for Informix
one year ago.

> Do you know how an RDBMS (like informix or Oracle or SQLServer) wiith an
> index on an attribute compares in its search completion time to your
> C-ISAM search with an index on that same attribute?

I can't compare it seriously - the Informix database contained more entries
but each entry has had only 3 attributes.
Example: A search for "surname=XXX" (base /, scope subtree, attribute to
return: telephone number) takes 0.3 seconds both for the X.500 server and
Informix 7.0. (2 processors, 256 MB, ...)

> In other words, does x.500 or LDAP have any deleterious effects on the
> search retrieval or does the speed of the RDBMS pass straight through?

A good server implementation can take advantages of the tree structure - so
it could (should?) be even faster.

> And finally, does the x.500 server you use create the tables in your
> RDBMS or do you need to map the DIT yourself to the RDBMS?

The server (DSA) creates the C-ISAM structure. There is no way to access
the C-ISAM database directly - you don't know how the DSA stores the data.
So the way how to get the data in is: load them via a tool. The tool I'm using
is a DAP client which adds all entries via DAP. (Similar to ldap_add.)
Speed is acceptable - first 40000 entries were loaded in one hour.
(And the only way to read the data is via DAP/DSP/LDAP.)