On 11 okt 2011, at 22:53, Ross Callon wrote:
I didn't mean to say that the IETF in general allows multiple solutions but
I think
it is accurate to say that the IETF has a less than 100% success rate of
preventing
multiple solutions.
Correct. We are not perfect.
I had four "proposals" for chat protocols when I was Apps AD. One of them was
NOT XMPP which now later seems to be what people use.
My take is that:
- If there is one proposal for a standard, it is "enough" for that proposal to
be "technically sound"
- If there are two proposals (or more), i.e. a situation where the market is to
choose between the multiple proposals that more or less solve the same problem,
there is an additional constraint, and that is that the proposals do not
interfere with each other
So if one have more than one proposal on the table that all move forward, they
must be technically sound AND also not interfere with each other.
This 2nd requirement is something that is not as easy to resolve as people
might think.
And specifically (we see in the MPLS case), the "bandwidth" of the liaison
connection between SDOs is not high enough to guarantee this. We learned that
the hard way between W3C and IETF.
Because of that the liaison coordination is only to resolve what SDO is
managing the multiple proposals. It can not handle the synchronization between
proposals.
And this to me is a key issue the whole MPLS discussion. It is just plain wrong
to try to manage all variants of multiple flavors of ice-cream across SDOs. And
why I strongly have the view I have on what has gone wrong.
Patrik
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf