Buzz roundtable: Directories, 09/27/99

Directories

Network World Fusion, 09/27/99

It sounds like there was some agreement that directories would play a key role here, and that's another topic I wanted to hit on. It seems like there are significant hurdles to overcome with respect to directories – the whole idea of integrating all your existing directories, internal politics over who controls what, just in general getting your directory house in order. What steps do you think users should be taking with respect to directories?

Kearns: You have to draw a line between the so-called meta-directories and what I like to call virtual directories. Meta-directories of the type Zoomit sells represent the master directory model. That's as opposed to the NetVision Synchronicity virtual directory model, where you use the directories you have, you simply connect them in some way to data that resides in other directories. In that case, there is no master model, there is no single point of control of the entire directory space. Whoever owns each directory continues to own it, you're just sharing information back and forth.

Nolle: That seems to be, to some degree, dodging the benefit along with dodging the issue. If independent ownership of directories is maintained, then even providing a technical mechanism to ensure the synchronization of the information content wouldn't necessarily provide any practical synchronization because the owners of the directories might not elect to operate that way. To a degree, if one assumes that we're going to move into some form of integrated directory environment, then at least part of the benefit is associated with the assumption that it's going to eliminate the inconsistencies among directories that tend to plague operations today.

Bradner [ON META-DIRECTORIES]: That would demand reengineering to a very large extent in many organizations, to an extent that I would find to be implausible.

Nolle: I agree with that. That really goes back to the higher level question of policy management and directory enabled networking and all of the other buzz words that are associated with these directory functions. We are confronted by the problem of either doing next to nothing and getting it by the current user mindset, or biting off a reasonable and justifiable chunk and colliding with so many existing implementations, political domains and other factors, that the implementation of it becomes implausible. So this is a technical remedy that's probably useful, but I don't know that we can really expect anything of it in the near term because the impact on the organization is too great.

Bradner: That's barking up the right tree. Having had to worry about some of this at Harvard, it's not something that I'm looking forward to getting too much further into.

Kearns: You have to decide which way you want to go. Do you want to change the culture of the organization and wipe out all those political differences…..

Bradner: ... You mean the ones that have taken 370 years to develop?

Kearns: Exactly. Or, do you want to find a way to work within that culture by allowing people to keep ownership of that data and yet still ensure that the data is replicated?

Bradner: I don't have any choice but the latter.

Kearns: Good. Then virtual directories are the way to go.

More of the roundtable