[RRG] Migration to IPv6-only addresses

[Robin Whittle <rw@firstpr.com.au>]

In the thread "End user network size [ [Q] draft-farinacci-lisp:
IPv4 address depletion]", Iljitsch van Beijnum wrote of two classes
of applications which present the greatest difficulties in a user
being happy with an IPv6-only IP address.

I initially wrote of the end-user networks having IPv6-only
addresses, but Iljitsch tended to discuss end-user computers having
IPv6-only addresses, with some servers in their network (or perhaps
the ISP network) having IPv4 addresses as well.

I found this analysis interesting, but there is so much devil in the
detail of all the forms of Internet communication, that I imagine
there are many other types of application which also present
problems, and which are not amenable to proxying.  For instance,
computer gaming applications - where real-time, direct, unmediated
communication is required, and where there is no general protocol -
just a bunch of separate protocols written by generally commercial
developers for mass-market users with Windows machines.

I am not convinced that IPv6 is simpler than IPv4, but even if it
was, the question is who is going to be happy with an IPv6 only
address in a situation where many other Internet computers (or as
few as 1% or less) have only IPv4 addresses.

Iljitsch, you wrote, in part:

> Fortunately, the 99% issue is only applicable to a small subset of
> all applications. IPv6 deployment issues depend on the type of
> application: client/server or peer-to-peer, and one-to-few or
> one-to-many.
>
> Client/server one-to-few would be email: you only talk to a very
> small number of mail servers.

I understand you mean the desktop machine only has an IPv6 address
and its email client software communicates using SMTP, IMAP and POP3
to the mail server of the network of which it is a part.

This assumes that a visiting person doesn't want to use the computer
for IMAP etc. to their own (IPv4-only) server.

I am keen to keep the MTA (Postfix) and IMAP server (actually
Courier Maildrop filtering, various spam and antivirus filters and
Courier IMAPD) on the same machine, within my network.  The MTA
(Postfix) certainly needs to have an IPv4 address in order to
interoperate with all other MTAs in the world.

I imagine that many other end-user networks insist on running their
own mail server, rather than paying the ISP to do so.  Years of mail
archives are huge, and there's no way an ISP wants to be responsible
for maintaining all that stuff.  For fast access, the IMAP server
needs to be on the local network.   If it is an IPv6-only network,
then it will be on a machine without an IPv4 address, but the MTA
needs to have an IPv4 address.  I guess the MTA could be separate -
at the ISP for instance, but I would much rather run the whole lot
on one machine, or on separate machines within my own network.


> If those servers that you use are dual stack, you can run
> IPv6-only on your workstation and still read and write mail
> without trouble.

I agree.  However, if the whole network has only IPv6 addresses,
which is a more specific form of the original question, then the
mail server can't be dual stack without tunneling to some network
which gives it a stable IPv4 address.


> Client/server one-to-many is web: you talk to many different web
> servers. You can only be IPv6-only once ALL these servers are dual
> stack.

Yes - we return to this below.


> Peer-to-peer one-to-few is BitTorrent: you download pieces of
> files from different peers, but you don't really care from which
> peer. So as long as the tracker (coordinating server) and a
> reasonable subset of all peers is dual stack, you're in business.

I agree in principle, assuming the BitTorrent system - perhaps just
the coordinating server system - is made to work with IPv6.
However, there are many such systems beyond BitTorrent.  Why would
the authors of BitTorrent et al. complexify their systems
appreciably (a lot more work, documentation, configuration items and
bugs) for IPv6 until there are an appreciable number of end-users
with IPv6-only addresses?

  (There is apparently an IPv6 client and tracker, after 2 days
   effort, at:  http://reboot.animeirc.de/bittorrent/ .)


> Peer-to-peer one-to-many is VoIP: in theory, you can dial any
> number, so you can only run IPv6-only once your SIP server and all
> possible peers are dual stack.

> Turns out you can easily move an application from the one-to-many
> to the one-to-few category by implementing proxies. So once your
> ISP (or IT department) has set up a dual stack HTTP/HTTPS proxy
> you can reach the web (in theory everything that uses TCP with the
> HTTPS proxy) and if your VoIP provider has an IPv6-capable gateway
> towards other networks, you can make calls over IPv6, too.

I am sure there are lots of applications where, due to performance
degradation, proxy load problems, the need for privacy, security and
encryption etc. it is impossible or very much second-best approach
to use an IPv6 - IPv4 proxy.

How many such applications are built to work with proxies?
BitTorrent isn't AFAIK.  I don't know a way of finding out 99% of
the applications people are actually using.  I think that any
broad-brush statement about whole classes of applications being
suitable for proxying would not withstand close scrutiny.


> So in an environment where only a subset of all possible
> applications must be supported (i.e., an enterprise network) it's
> entirely possible to ditch IPv4 and proxy to the IPv4 world at the
> edge of the network.

I think this may well be the case in highly controlled enterprise
networks such as banks.  But those networks are doing fine with NAT
and IPv4 anyway - probably having thousands of users accessing the
Net via a handful of IPv4 addresses.  So shortage of IPv4 address
space is not going to be an impetus to IPv6 adoption, especially if
and when LISP etc. is available to provide stable PI IPv4 addresses
to such networks in smaller than 256 increments.


>> I think we are still going to be clamouring for IPv4 addresses in
>> ten or fifteen years time.  Maybe we will be stuck with it
>> forever.  I hope a new routing and addressing architecture can be
>> implemented around 2012.
>
> I find it strange how people can make predictions that far in the
> future. Less than 1% of all people that are now running IPv4 were
> doing so 15 years ago. If you can go from 0 to 4 in 15 years, why
> not from 4 to 6 in as many or fewer years?

15 years ago, there was no installed base.  IPv4 was not
incompatible with existing practice.

IPv6-only is incompatible with existing IPv4 practice and provides
no obvious benefit for a dual-stack machine - until perhaps there
are large numbers of IPv6-only users.  In that case, if there is a
performance improvement, it is only to the extent that IPv6-only
computers suffer a performance problem or communications
impossibility when trying to communicate with an IPv4-only machine,
which is what I think must be prevented before IPv6 will ever be
widely adopted.

If MS-DOS and IBM PCs hadn't been widely deployed when the Mac was
introduced, everyone would probably be running Macintosh software
now and all PCs would probably be based on 68000 derivatives.
(While I don't use or like the Mac, I think this would have been a
much clearer path - starting with a decent 32 bit CPU.)

The IBM 360 was introduced in 1964 and I understand that in the 43
years since (the 360 was produced until 1977), all IBM mainframes
have been compatible with this now completely antiquated
architecture - with no stack pointer and with software based on
EBCDIC (according to Wikipedia, also for backwards compatibility).

IPv6 is not backwards compatible with IPv4 - due to the
impossibility of any larger addressing system and/or set of IP
protocols (as far as I know) being backwards compatible with
installed IPv4 operating systems and applications.

I can't avoid the conclusion that we are likely to be stuck with
IPv4 and NAT forever.  I would be happy for someone to prove me wrong!

  - Robin



--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg