Re: [Int-area] Fw: Continuing IPv10 I-D discussion.

joel jaeggli <joelja@bogus.com> Sat, 01 April 2017 05:11 UTC

Return-Path: <joelja@bogus.com>
X-Original-To: int-area@ietfa.amsl.com
Delivered-To: int-area@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id CE266128CFF for <int-area@ietfa.amsl.com>; Fri, 31 Mar 2017 22:11:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.901
X-Spam-Level:
X-Spam-Status: No, score=-6.901 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_HI=-5, RP_MATCHES_RCVD=-0.001] autolearn=ham autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZS9E9e2DQOTJ for <int-area@ietfa.amsl.com>; Fri, 31 Mar 2017 22:11:23 -0700 (PDT)
Received: from nagasaki.bogus.com (nagasaki.bogus.com [IPv6:2001:418:1::81]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 50998127275 for <int-area@ietf.org>; Fri, 31 Mar 2017 22:11:23 -0700 (PDT)
Received: from mb.local ([75.104.70.255]) (authenticated bits=0) by nagasaki.bogus.com (8.15.2/8.15.2) with ESMTPSA id v315B8vu081866 (version=TLSv1.2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128 verify=NOT); Sat, 1 Apr 2017 05:11:16 GMT (envelope-from joelja@bogus.com)
X-Authentication-Warning: nagasaki.bogus.com: Host [75.104.70.255] claimed to be mb.local
To: Khaled Omar <eng.khaled.omar@hotmail.com>, Tom Herbert <tom@herbertland.com>
References: <D502B93A.74992%lee@asgard.org> <AM4PR0401MB224189BDD22CD327CF280AA3BD340@AM4PR0401MB2241.eurprd04.prod.outlook.com> <alpine.DEB.2.02.1703310806130.30226@uplift.swm.pp.se> <1d67d033-8a0f-c7eb-ae37-ec99f5a34660@kit.edu> <AM4PR0401MB2241FCE296DCD88D7D065520BD370@AM4PR0401MB2241.eurprd04.prod.outlook.com> <CALx6S35umePtAb-noP_CiXOh9Kf8j00oCVPSevci6EE9fyxTqQ@mail.gmail.com> <AM4PR0401MB2241D2A7F373B24116E999BCBD370@AM4PR0401MB2241.eurprd04.prod.outlook.com>
Cc: "int-area@ietf.org" <int-area@ietf.org>
From: joel jaeggli <joelja@bogus.com>
Message-ID: <5fbcf078-4e05-3ea8-01a8-e0228659eb86@bogus.com>
Date: Sat, 1 Apr 2017 00:11:01 -0500
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:45.0) Gecko/20100101 Thunderbird/45.8.0
MIME-Version: 1.0
In-Reply-To: <AM4PR0401MB2241D2A7F373B24116E999BCBD370@AM4PR0401MB2241.eurprd04.prod.outlook.com>
Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="288MDKoWAtForrnFeQmT7AmBUkLBaKdrO"
Archived-At: <https://mailarchive.ietf.org/arch/msg/int-area/iG2bDEARwAs5kZB-CtyQ5WLujeU>
Subject: Re: [Int-area] Fw: Continuing IPv10 I-D discussion.
X-BeenThere: int-area@ietf.org
X-Mailman-Version: 2.1.22
Precedence: list
List-Id: IETF Internet Area Mailing List <int-area.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/int-area>, <mailto:int-area-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/int-area/>
List-Post: <mailto:int-area@ietf.org>
List-Help: <mailto:int-area-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/int-area>, <mailto:int-area-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 01 Apr 2017 05:11:26 -0000

On 3/31/17 09:24, Khaled Omar wrote:
>> "all OSs will be updated to support IPv10 which is an easy
>> task"...
>>> What makes you think it is ever an easy task to get all OSes to
>>> uniformly support anything? Please provide an implementation so
>>> we can evaluate the prospects.
> 
> For the 2nd time:
> 
> "I shouldn't list all devices that need to support IPv10, simply,
> anything will process a L3 packet, should understand that the IPv10
> packet can contain a mixture of IPv4 and IPv6 addresses."

Assuming for the sake of argument that I am designing a forwarding asic.
When I pass over the ethertype code in the frame, I decide what
forwarding lookup path I am going to exercise the ipv4 or ipv6 one. If i
choose door number 1 the ipv4 path  due to 0x0800  and destination mac
being me the byte offset for the source is 12 and the destination is 4
bytes later at byte offset 16. if these fields are of variable length
then a 16 byte source address overflows the fixed offset  where the
destination  is supposed and the destination address is lost to me
because it's where optional fields might otherwise be.

If I choose the v6 path due to 0x86DD then I find the source address
address at byte 8 and the destination beginning at byte 24 there, there
are of course several ways to represent a an ipv4 mapped address within
ipv6, none of them are of any use to an native ipv4 speaker.

The fields in the packet header are not self-described (they could be,
packet header could be json for example  but in the interest of
performance they were / are not), therefore  in the interestes of
clarity a new ethertype for ipv10 with either a fixed size source and
destination address size  or a whay of specifying the offset, and a bit
or multiple bit field for the AF of both the source and destination
would greatly facilitate the design of a parser that would be able to
handle the selection of AF for lookup. at the point you've go an
entirely new header which is not compatible with either V4 or V6
parsers. in practical terms that means the useful deployment of a v10
compatible parser at internet scale is approximately as hard as
deploying the the v6 compatible parser would be if we started now
instead of in 1995 when the internet was much smaller.

joel

> 
> -----Original Message----- From: Tom Herbert
> [mailto:tom@herbertland.com] Sent: Friday, March 31, 2017 4:19 PM To:
> Khaled Omar Cc: Bless, Roland (TM); int-area@ietf.org Subject: Re:
> [Int-area] Fw: Continuing IPv10 I-D discussion.
> 
> On Fri, Mar 31, 2017 at 5:13 AM, Khaled Omar
> <eng.khaled.omar@hotmail.com> wrote:
>>> As has been stated again and again. Your proposal would have been
>>> interesting if it was presented in 1995, or perhaps even in
>>> 2000.
>> 
>> FYI, IPv10 will allow IPv4 to communicate to IPv6 and vice versa,
>> how can it be interesting if it was presented before IPv6 was even
>> developed !
>> 
>>> Most likely, even if Microsoft could be convinced that IPv10 is
>>> something they need to support, this would only happen in Windows
>>> 10. Then we have the rest of the ecosystem with access routers,
>>> load balancers, SAVI-functionality for BCP38 compliance in access
>>> devices, core routers etc.
>> 
>> Please, let's not be against ourselves, all OSs will be updated to
>> support IPv10 which is an easy task, OSs will not require support
>> for a new IP version like IPv6, they will just be enabled to
>> support the encapsulation of both version on the same L3 packet
>> header.
> 
> "all OSs will be updated to support IPv10 which is an easy task"... 
> What makes you think it is ever an easy task to get all OSes to
> uniformly support anything? Please provide an implementation so we
> can evaluate the prospects.
> 
> Tom
> 
>> 
>> Also, networking devices will be upgraded to understand the new
>> IPv10 packet, I said earlier I don't mind If the process will take
>> some time but we should eventually reach consensus, I shouldn't
>> list all devices that need to support IPv10, simply, anything will
>> process a L3 packet, should understand that the IPv10 packet can
>> contain a mixture of IPv4 and IPv6 addresses.
>> 
>> 
>> -----Original Message----- From: Int-area
>> [mailto:int-area-bounces@ietf.org] On Behalf Of Bless, Roland (TM) 
>> Sent: Friday, March 31, 2017 9:51 AM To: Mikael Abrahamsson Cc:
>> int-area Subject: Re: [Int-area] Fw: Continuing IPv10 I-D
>> discussion.
>> 
>> Hi Mikael,
>> 
>> thanks for clarifying again, everything +1!
>> 
>> Regards, Roland
>> 
>> Am 31.03.2017 um 08:17 schrieb Mikael Abrahamsson:
>>> On Thu, 30 Mar 2017, Khaled Omar wrote:
>>> 
>>>> You can read the IPv10 I-D again and all your concerns will be
>>>>  obvious, I don't mind if you have already a series of new
>>>> questions that will add a new value to the discussion but the
>>>> time to deploy IPv10 is an important factor.
>>>> 
>>>> We need consensus after understanding how IPv10 works and how
>>>> it will be deployed.
>>> 
>>> As has been stated again and again. Your proposal would have been
>>>  interesting if it was presented in 1995, or perhaps even in
>>> 2000.
>>> 
>>> Let me give you an IPv6 deployment timeline:
>>> 
>>> Standards were worked out in the mid 90-ties, afterwards
>>> operating system vendors started working on it and "real" support
>>> started cropping up in the early to mid 2000:nds, with a large
>>> milestone being Windows Vista in 2006, where as far as I know
>>> this was the first widely used consumer operating system to
>>> implement this. It then took until Windows 7 timeframe around
>>> 2010 before people started moving off of Windows XP in ernest,
>>> and we're still seeing Windows XP in non-trivial numbers. So now
>>> in 2017 we're seeing most operating systems have comprehensive 
>>> (albeit perhaps not as well-tested as we would like) support for
>>>  IPv6, where the application ecosystem still has a way to go.
>>> We're still working on better APIs to handle the dual-stackedness
>>> problem.
>>> 
>>> Most likely, even if Microsoft could be convinced that IPv10 is 
>>> something they need to support, this would only happen in Windows
>>> 10. Then we have the rest of the ecosystem with access routers,
>>> load balancers, SAVI-functionality for BCP38 compliance in access
>>> devices, core routers etc. Most of these will require a hardware
>>> fork-lift in order to support your proposal, because they do not
>>> forward packets in a CPU, they forward it in purpose-designed
>>> hardware that is a lot less flexible in what they can do.
>>> 
>>> So even if we all united now (which won't happen) around your
>>> IPv10 proposal, it would take 5-10 years before the first devices
>>> out on the market had support for it. Probably 5-10 years after
>>> that before support is widely available.
>>> 
>>> IPv10 would delay and confuse deployment of something that is not
>>> IPv4. While IPv6 is not perfect, there are now hundreds of
>>> millions of devices on the Internet with IPv6 access. It's proven
>>> to work, it's not perfect, but we have a decently good idea what
>>> to do to make it better.
>>> 
>>> IPv10 is only injecting FUD into where we need to go debate,
>>> which is IPv6 deployment for all.
>>> 
>>> Please stop.
>> 
>> _______________________________________________ Int-area mailing
>> list Int-area@ietf.org 
>> https://www.ietf.org/mailman/listinfo/int-area
>> 
>> _______________________________________________ Int-area mailing
>> list Int-area@ietf.org 
>> https://www.ietf.org/mailman/listinfo/int-area
> _______________________________________________ Int-area mailing
> list Int-area@ietf.org 
> https://www.ietf.org/mailman/listinfo/int-area
>