Leonir Hoxha wrote:
Hi Khalid,
Are you aware what you are proposing!!?
Those RFCs are incorporated into all the Vendors for many years now.
Hundreds of Experts have been working on these Protocols, and you just
want them to obsolete?
What is your value proposition, I have read your letter, but honestly it
doesn't give me any feeling that it is a real solution.
We do have IPv6/IPv4 already talking to each other, why do you think an
additional Address Family would make our life easier?
Please consider this, and try to answer these questions to yourself.
And last, please STOP sending mails to ietf(_at_)ietf(_dot_)org, this is not
the right way to do it.
On the contrary, the calls for someone to stop sending work that is outside the
scope of any existing working group to this list are "not the right way to do
it". If you believe this belongs in an existing group, redirect as appropriate.
If this is truly a new protocol and doesn't fit alongside existing work, it
belongs here to help determine consensus about forming a new working group.
That said, at best this belongs in v6ops as a transition approach. In reality,
it is inconsistent with itself to the point where it can't be implemented, even
if you wanted to. The section 3 descriptions use variable length headers, while
the section 4 description of the proposed "new protocol" header is fixed
length. Independent of that, the author has not recognized existing technology
(RFC 4921 Sections 2.5.5.1 & 2.5.5.2) that does more than what this draft is
proposing despite being pointed at that a month ago.
Finally, this proposal does nothing to solve the problem it identifies, legacy
IPv4 hosts in the enterprise environment that will not migrate. There appears
to be an unstated assumption that administrators of legacy hosts will make the
changes necessary for this inconsistent and underspecified proposal, despite
the demonstrated fact that they are unwilling to make the well documented
changes to deploy IPv6 because they simply refuse to make a change, or to
learn something new.
In the beta version of Windows XP I made sure IPv6 was ON-BY-DEFAULT, with no
knob to turn it off, because I knew that given the option enterprise
administrators would turn it off. I lost the battle in the shipping version
though, and sure enough the first things that showed up on search engines were
instructions for how to turn IPv6 off. The lack of deployed end systems limited
the willingness of app developers to bother making their changes, which in turn
fed the resistance to change by the intermediate network administrators. Fast
forward to the point where IPv6 is required, and forcing the changes outward
from the middle is proving to be as difficult as predicted. The IPv10 proposal
does nothing to solve this fundamental problem, and actually amplifies it by
removing what clarity there is in dual-stack network administration and
diagnostics.
Tony