Newbie : DFS : Why a physical root is required?

[Guest redheart27@rediffmail.com]

I have a fundamental question that is not directly answered anywhere,

hence I seek your help. I wish to understand the idea behind the

design. My intention is to learn.

 

This is how I (as a newbie) would expect DFS to be designed. (Assume a

Domain environment)

A DFS service runs on the DC/member servers. The namespace and other

configuration is replicated among these servers. The service will

listen to requests from clients and answer them by looking up its

data.

 

(a) As per the MS implementation, an empty physical directory

structure is created to represent the namespace. Why was this unusual

data representation chosen?

 

(b) Why does Active Directory(AD) duplicate the information of the DFS

root? In fact, why does AD get involved at all, except perhaps, to

provide the information for clients to locate the DFS servers?

 

Please help. Thanks.

[Guest Juergen Kluth]

Re: Newbie : DFS : Why a physical root is required?

 

HI,

> (a) As per the MS implementation, an empty physical directory

> structure is created to represent the namespace. Why was this unusual

> data representation chosen?

 

Otherwise the namespace was just the namespace, and you couldnt do anything

with it.

 

You cant make coffee on it, nor can you save files on it.

I'm sure you have no namespace, you can make coffee on it,

even if it was a java namespace, do you?

 

May be you could write the namespace down on paper.

Then for sure, one comes and asks, Why the hell did he involve paper with it

?

> (b) Why does Active Directory(AD) duplicate the information of the DFS

> root? In fact, why does AD get involved at all, except perhaps, to

> provide the information for clients to locate the DFS servers?

 

Isnt it a nice idea ? Saving you from running arround in your (later)

company,

making miles and miles every day with your cheat of paper holding your

namespace,

duplicating it to the other employees!

 

;-)

JK