Re: [v4tov6transition] draft-arkko-ipv6-transition-guidelines WGLC

[Cameron Byrne <cb.list6@gmail.com>]

2010/8/28 YangGL <iamyanggl@gmail.com>:
> Tests in my lab have proved that many popular applications cannot work on
> IPv6-only network with NAT64, such as IM, P2P, games, and part of video. WEB
> and part of mail (Outlook and Outlook express) are the only applications we
> can find working properly with NAT64. But there are more than 50% traffic is
> P2P, WEB traffic is less than 20% on CT’s network. I think it is not a good
> news to NAT64.
>

I agree, NAT64 is not great for fixed line operators today, but i am
sure someone has suggested DS-lite to you.

Mobile wireless in the USA is a different case, 90+% of my traffic is
Web (tcp 80 and 443) and Email (pop,smtp, imap ...).  And, 40+% of my
overall aggregate ISP bandwidth goes to content providers that have
substantial capability to deliver IPv6 content today (Google,
Facebook, ...).  As draft-arkko-ipv6-only-experience states, mobile
can reach near 100% functionality today with IPv6-only on the
handsets.  The expedience varies greatly between handset manufactures,
but the gap is closing quickly.  Most folks know Nokia is a leader in
IPv6, and now we see substantial IPv6 efforts from Android and Apple
(Windows phone 7?).

From an access network perspective, i have found it is extremely
important to partner with content providers as well as applications
developers and have an open dialog about native IPv6 as well as how
transition mechanisms will impact the user experience.  In my real
world experience, the content providers take user experience very
seriously, and they know CGN is harmful and IPv6 is a near term
reality.  And, i have taken the wiki approach to fix IPv4 literals
http://groups.google.com/group/ipv4literals , at least the ones that
can be fixed as i encounter them in my real world pilot network.

I believe the point here, as someone else likely said, there is no one
size fits all.  Using a search engine, you can find enough general and
enough specific information to find what works best for you.  Review
it with your vendors, do labs, trials, pilots, betas .... learn,
repeat, deploy at small scale, learn, then deploy at large scale,
monitor, learn, fix ... probably need to do training in there
somewhere too.  From a technology evolution and process perspective,
there is nothing new with IPv6.

Finally, it is interesting you said that 50% of the traffic is P2P and
you are concerned P2P will not work with IPv6.  According to this
link, P2P traffic is driving IPv6 growth http://tinyurl.com/lkjwhq and
http://tinyurl.com/25ka5jp  From the operators that i have talked to,
rolling out IPv6 natively in the fixed-line environment reveals a
substantial amount of P2P  IPv6 traffic.  uTorrent is an IPv6
vanguard.  Others will follow if they have not already in making their
apps work with IPv6.

Once again, the knife cuts both ways. Some P2P applications will
thrive in IPv6 (seems obvious given e2e nature of things) while others
will be forced to evolve or be deprecated by their users.

Regards,

Cameron
=======
http://groups.google.com/group/tmoipv6beta
=======