Kelly, Terence P wrote:
I'm a researcher and one of my interests is in the trade-offs between performance and other properties (e.g., fault isolation/security) that various software architectures offer.
Furthermore I want to include OpenLDAP in an apples-to-apples comparison involving other applications with multi-tier architectures in which the various tiers run on separate physical machines.
Therefore I'm interested in any way of running the underlying database on a different physical machine then the slapd application, even if the performance overhead is substantial.
Does that make sense?
Thanks in advance for any suggestions you can offer.
It doesn't sound like an apples-to-apples comparison is really possible, since the primary mode of operation for the OpenLDAP server is monolithic.
However, you can of course configure OpenLDAP's back-sql to talk to a remote RDBMS server if that's what you really want to do. I don't consider this a realistic use of OpenLDAP, but the support exists for those masochistic enough to go there.
On the other hand, if what you're really testing is multi-tier from the client perspective, then it should suffice to call the LDAP client the first tier, and a remote OpenLDAP server the second tier. If you want to introduce more levels of indirection you can simply stack back-ldap proxies between the client and the ultimate data server, but again, adding layers for the sake of layering seems pretty unrealistic to me.