Scaleability?

[Dan Armbrust <da...@gmail.com>]

We currently use OpenLDAP for storing large hierarchical databases of 
medical terminologies as part of our LexGrid project 
(http://informatics.mayo.edu/LexGrid/)- but would welcome an alternative 
to OpenLDAP for numerous reasons that I won't go into.

I have just started exploring your directory project, and I am trying to 
get an idea of what its current capabilities and limitations are. 

I realize that I could probably answer most of these myself by just 
trying it... but I'll take the easier route (for me) first and just ask :)

Does the ldap server support custom schemas?
Can you load data into the server that doesn't conform to the base schema?
What is the saleability like with the current backend?  Is it capable of 
working with 10,000 entries, 100,000 entries, millions?
Is the full ldap search syntax supported?
Including wildcard searches? 
Can it be loaded from an LDIF file?
Is there an FAQ page for your project on the website or the WIKI where I 
should be asking these questions?

Thanks,

Dan

Re: Scaleability?

[Alex Karasulu <ao...@bellsouth.net>]

Dan Armbrust wrote:

> We currently use OpenLDAP for storing large hierarchical databases of 
> medical terminologies as part of our LexGrid project 
> (http://informatics.mayo.edu/LexGrid/)- but would welcome an 
> alternative to OpenLDAP for numerous reasons that I won't go into.
>
> I have just started exploring your directory project, and I am trying 
> to get an idea of what its current capabilities and limitations are.
> I realize that I could probably answer most of these myself by just 
> trying it... but I'll take the easier route (for me) first and just 
> ask :)
>
> Does the ldap server support custom schemas?

Yes but you cannot dynamically add them.  OpenLDAP has the same limitation.

> Can you load data into the server that doesn't conform to the base schema?

Sure right now there is no strict schema checking and when added (easy 
to do) schema checking can be turned off.

> What is the saleability like with the current backend?  Is it capable 
> of working with 10,000 entries, 100,000 entries, millions?

The current backend should be able to handle up to 1 million entries 
easily but may need indices to be rebuilt to keep up performance if the 
btrees are no longer balanced.  JDBM is a few Java DBM implementation.  
Basically very similar to SleepyCats JE or Berkeley DB.

> Is the full ldap search syntax supported?

No we do not support approximate matches with soundex algorithms.  
Basically a ~= based assertion is interpretted as an equals assertion.  
Why?  I never really saw it work right or no one really makes canned 
searches depend on this assertion operator.  It's usually adhoc searches 
that make use of it if at all.  Plus it complicates building different 
kinds of indices for approximate matching.  Not worth the trouble it 
causes but all else is there.

>
> Including wildcard searches? 

Yep!

> Can it be loaded from an LDIF file?

The search?

>
> Is there an FAQ page for your project on the website or the WIKI where 
> I should be asking these questions?

Take a look at the user guide and other documents here:

http://directory.apache.org/subprojects/apacheds/index.html

-Alex