Re: Comments on draft-ietf-multi6-v4-multihoming-02

On 18-nov-04, at 20:49, Joe Touch wrote:

> | This is exactly how we always get into trouble: add
> | protocols that break the assumptions of previous protocols. It would 
> be
> | much better to change TCP so it has all the information and can make 
> the
> | right decisions.

> Er, well, that would argue that TCP should do the multihoming too, not 
> a
> shim layer. We already discussed why _that_ isn't the case. For similar
> reasons, this is true here too.

I disagree that these cases are similar enough that the tradeoffs for 
one can be applied to the other.

> | TCP is _NOT_ the muxing layer. IP is.

If memory serves me, port numbers happen in TCP and UDP, not IP... 
(Which is a mistake anyway, IMNSHO.)

But anyway, we have a lot of other bridges to jump off before we reach 
this one, so we can hash this out at some point in the future in the 
then appropriate forum.

> Check the IETF proceedings for webmux, HTTP-NG, and SOAP.

I will, thanks.