I've been investigating LDAP for a couple of weeks now and I DON'T GET IT. Perhaps I'm missing something, but why all the complexity when using openLDAP etc when I could just define my own DB schemas without the extra layers of abstraction? Can anyone please tell me what I'm missing.
Basically all I'm looking to do is lookup objects not people etc on our internal work wouldn't I make my life infinitely simpler writing simple code that accesses a DB or file anyway(just like LDAP does) and customise it to fulfill the cryptic LDAP schema spec?????
Performance issues aside - by all means - if you are more comfortable writing your code to query a static file or database - go ahead.
Not sure what you mean about LDAP being a file - but the reasons most folks use LDAP are performance, accessibility utilizing LDAP over TCP/IP, and security.
From what I gather you would be better off sticking with something you know.
Thanks Chris I appreciate the response. I have subsequently got everything up and running including reference addresses for javaObjects etc. My thinking still is however why not just have schema files which convert to db entries using dsml or similar to make sure all input is valid. Then have all these entries from the various tables based on the individual LDAPV3 compliant schemas read into a value object/hashmap or similar at startup. Any realtime updates could immediately repopulate the valu object and so on. Performance would be better and the system would still be LDAPv3 compliant as it is still based on compliant schemas.
Almost all the commercial offerings I'm currently investigating do not use LDAP for anything other than user authentication and are not nearly taking advantage of the power of LDAP as we know it. In short they are mostly lip service.