Re: [dhcwg] [Softwires] We can change the world in a 1000 ways (IPv4 over IPv6)

[Mark Townsley <[e-mail]>]

On Nov 13, 2013, at 5:05 PM, Ted Lemon wrote:

> On Nov 13, 2013, at 10:49 AM, Ole Troan <[e-mail]> wrote:
>> is there a problem here, or should we just accept that sometimes the IETF
>> will generate ten sets of publications solving more or less the same problem?
> 
> If I'd been area director earlier in the process

You rang?

> I might have just shut the working group when it became clear that the principals couldn't agree on a proposal, and required that they come to agreement before a BoF would be approved.

Not chartering work doesn't always help as it can scatter people to the winds to create their own parallel efforts. The 13+ IPv4 over IPv6 solutions you see were born from the rubble of trying to keep DS-Lite as the only chartered work softwires would look at for IPv4 over IPv6 with some kind of built-in IPv4 address sharing.

Ted, when we started softwires, I noted very clearly that I would reject every transition protocol to come my way until the WG agreed on a finite set of problem-spaces and associated solutions. Note I didn't say reject _chartering_ the work or having a BoF, I said reject the proposals that did not come via the WG (e.g., as an end-run to the chartered effort). This gave people a common place to work, and at the start that's exactly what it did. We held 3 interim meetings, and killed many more proposals than we advanced. By the time my AD tenure was up, we had identified L2TP for point-to-point stateful tunnels from hosts and such (largely because it was already in so many places already). We had a second solution for larger core networks (essentially a generalization of MPLS' 6PE). The third was 6rd, building on 6to4 in a manner that was far more palatable to an operator to deploy and had the obvious scaling properties over L2TP.  So, the happy medium for what were deployed as largely IPv6 over IPv4 solutions ended up somewhere between "one-size-fits-all" and "everything-goes". 

To my memory the 4/6 mess kicked a bit later. The mess started not because the WG allowed too much in its charter, but because it was rejecting so much work outright that it didn't give people a trusted place to come together and work towards a compromise sooner rather than later. 

My point is that sometimes we fail when we don't give people the right environment to work together and make compromises early on in the hopes of rising the tide for everyone. If groups remain in their respective corners working on their own for too long, they are bound to create their own parallel paths. The longer that is allowed to happen, the more entrenched the solutions become, until the IETF has lost its ability to do anything constructive beyond publishing everything.

- Mark



>   But it's much too late in the process to do that now.   And I don't even know if that would have produced a better outcome.
> 
> I don't think we should accept that this has to happen every time, and I think we should try to prevent it happening in the future.   But there is no sense crying over spilt milk.
> 
> _______________________________________________
> Softwires mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/softwires