Short version: Tony wanted comments on his and Lixia's draft
recommendation. I am commenting (AKA discussing,
AKA debating) on some aspects of this, including
the applicability of ILNP to IPv4 - which I argue
needs to be mentioned explicitly since it has
significant practical difficulties in addition to
those which apply to ILNP for IPv6.
Ran and Tony have not yet responded.
Hi Ran,
What are the advantages of ILNP for IPv4 over hIPv4?
They both involve host changes and sending extra information with
every packet. If hIPv4 packets have a new IP header option, as you
propose for ILNP for IPv4, then both architectures will also require
upgrades to some, many or all DFZ and other routers.
Both architectures provide a larger (64 bit) Identifier space. I
recall Patrick is considering alternatives to header options. Could
you include the ILNP header and the rest of the packet within a UDP
header? I think that would get around the need for a new header
option and so for the need to upgrade DFZ and other routers.
You wrote:
> Robin's notes about ILNP and about what
> I have said/not said/written/not written
> continue to be overwhelmingly inaccurate.
This is an unsupported assertion.
You spent significant bandwidth accusing me of having deficient
understanding of ILNP and scalable routing in general - when you
could have used that bandwidth to explain to me and others how ILNP
would work for IPv4. If you criticise people on the list they are
likely to defend themselves.
> I've been asked not to respond to his notes,
> on the reasonable basis that nothing in
> this thread is relevant to producing the
> RRG document.
Ah - so there are other people who don't like debate.
I was trying to have a meaningful discussion (AKA debate) about your
proposal. See:
http://www.ietf.org/mail-archive/web/rrg/current/msg06578.html
At first you refused to discuss your proposal - derided my lack of
understanding. Now that I have a better understanding, you still
don't want to discuss your proposal.
I am happy to discuss/debate my proposal with anyone. If I think
they don't understand it well, then I try to improve their
understanding. Quite likely I will learn something about scalable
routing, about my own proposal, or about how to explain it better.
I argue that the question of ILNP's applicability to IPv4 is
something Tony and Lixia's recommendation needs to be explicit about.
The current draft of the text:
http://www.ietf.org/mail-archive/web/rrg/current/msg06521.html
has nothing to alert the reader that ILNP is primarily designed for
IPv6, and that in the current header option proposal for IPv4 would
require upgrades to some, many or all DFZ routers before anyone could
use ILNP.
Tony's request regarding that draft was "Please comment." - which is
what I am doing.
- Robin
_______________________________________________
rrg mailing list
rrg@irtf.org
http://www.irtf.org/mailman/listinfo/rrg
