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

Re: search mechanism in openldap

On 6/1/06, Howard Chu <hyc@symas.com> wrote:

Kamal R. Prasad wrote:

thanks for the info.

Does openldap optimize searches in some way? Who exactly handles the
> searches when a client sends in a request? is it slapd or berkeley
> db4? What
> is the performance overhead of running slapd on the same box just to get
> this functionality?
Berkeley DB is obviously not a relational database, it knows nothing
about filters or query languages. It sounds to me like you haven't read
enough documentation yet to warrant one approach or another. If you're

I haven't read much about ldap -but the basic idea is not to provide an ldap
interface, but to store information, tag it and retreive it efficiently. I
am just trying to guage where slapd can do a better job than if we were to
write the code ourselves.

That aside, if I have multiple trees -each of which can lead to an object (
i.e. if more than one tag/filter can result in the same object), is the
right way to represent it an alias?
Are aliases well-supported?

Secondly, if I want to send a query to slapd (get me objects associated with
given tag aka part of tree X AND objects part of tree Y i.e. associated with
tag Y)

worried about performance overhead, the only thing to do is actually run
it and measure it to see if it meets your requirements.

-- Howard Chu
Chief Architect, Symas Corp.  http://www.symas.com
Director, Highland Sun        http://highlandsun.com/hyc
OpenLDAP Core Team            http://www.openldap.org/project/