[OpenAFS-devel] Re: Moving Forwards

[Derrick Brashear]

On Mon, Sep 10, 2012 at 2:29 PM, Andrew Deason <adeason@sinenomine.net> wrote:
> On Mon, 10 Sep 2012 13:03:03 -0500
> Troy Benjegerdes <hozer@hozed.org> wrote:
>
>> I understand rx with v6 is the first step. Let's imagine I have it.
>
> Well, I can 'imagine' it, but... people may take you more seriously if
> that were actually done. If you get that done (either by coding
> yourself, or just by paying someone to), it shows that progress can be
> made and that you're serious about it, etc etc.
>
>> What's next so I can get an isolated linux client, and isolated linux
>> server talking AFS over IPv6?
>
> Request RPC code points from the AFS-3 registrar, and say they are for
> institution-private use. You don't need to say anything else. I _think_
> the only thing you absolutely require are RPCs for the VL service.
> 'Normally' you would want some for the RXAFS (client<->server) and
> VOLSER (vos<->volserver, volserver<->volserver), but at least for a
> 1-server scenario, you don't need that.
>
> Then you (or someone you pay, etc) implement the client and server stuff
> for the RPCs you design. You don't need approval from anyone else. As
> long as you don't try to distribute the result as 'AFS' (or anything
> with AFS in the name), but rather either keep it completely private, or
> be clear that it is a non-standard extension to OpenAFS, I have trouble
> seeing you encountering trademark issues. Since... well, plenty of other
> similar modifications (unrelated to ipv6) already exist.

and figure out where you will store the addresses, since at least as
is the vldb won't.

if you fork, you could define your own format there and store it. all
your vldbs would have to be
"TFS" vldbs.




-- 
Derrick