[OpenAFS] MR-AFS

[Derrick J Brashear]

[You need not reply to me and the list: I'm on the list]

--On Thursday, March 15, 2001 09:14:41 PM -0500 Mitch Collinsworth 
<mitch@ccmr.cornell.edu> wrote:

>
> On Thu, 15 Mar 2001, Derrick J Brashear wrote:
>
>> Well, we can't just switch by default on existing platforms because it
>> would break compatibility, but it's a desired goal to have the NAMEI
>> fileserver working on all supported platforms, and for configure to have
>> an option to build that way (it's rather messy to build both a regular
>> and a namei fileserver, but it could be done; it's not vastly different
>> than the whole "t" versions for threaded environments. but... how many
>> variants do you build? where do you draw the line?)
>
> The whole new features/compatibility with IBM AFS  discussion is one
> that deserves careful consideration.  Of course we don't want OpenAFS
> to run off and make itself non-interoperable with IBM AFS, at least in
> the short run.  But IBM has not committed to releasing any new versions
> beyone 3.6.  Does this mean *all* future versions of OpenAFS will have
> to remain interoperable with IBM AFS 3.6?  Will it someday become a
> ball-and-chain that prevents adding useful new functionality?  Has any
> thought been given to maintaining compatibility in the short run but
> eventually giving up that requirement?  And if so how would such a
> decision be made?

Obviously at some point it will no longer be interesting to maintain 
compatibility with IBM, especially if they stop moving forward with the 
product. But for now, obviously, a large amount of the interest in OpenAFS 
will come from sites which already have some sort of deployed base, and so 
it makes sense to give them something they can use in the mix in their 
environments. I can only speak for me, but I think this is a "we'll cross 
that bridge when we come to it" situation. If AFS has a future it can 
clearly not be constrained by what it is now, but if not for what is it 
now, what is it, really?

-D