[arin-discuss] urgency of IPv6

[Ted Mittelstaedt]

On 6/28/2010 1:36 PM, Owen DeLong wrote:
>>>
>>> Who cares?  The important thing is that new eyeball users that are
>>> unable to get IPv4 addresses can get to the content without bizarre
>>> hacks to give them horribly degraded IPv4 connectivity.
>>>
>>
>> I don't get horribly degraded IPv4 connectivity when I surf the web from
>> my Windows Mobile 5 Smartphone on Sprint's network, and my wife doesn't
>> get horribly degraded connectivity when she surfs the web from her
>> Android phone on the T mobile network - but both those phones are on an
>> IPv6 network, using some bizarre IPv6-IPv4 proxy back at the cell companies NOC.
>>
>
> Right... Neither of you is a post-runout new eye-ball at this time and no, you are not
> correctly understanding how the cellular network you are using is actually working.
>
> First, neither of those networks is IPv6 yet, if you check, you'll see that your phones
> still just have IPv4 addresses.

While there is no way to go into any setting on the phone and check it's 
IP address, someone wrote a (free) diagnostic network app you can run on 
WM5 that DOES tell the actual number on the phone.  It IS an IPv6 
number.   Most people probably are confused because going to 
whatismyip.com or some such gives them an IPv4 address.

I'll check the Android phone but I strongly doubt, with their brand new
phone, that it uses IPv4.  I don't think they have enough numbers for
that, frankly.

> Eventually, as I understand the plans from both of
> those providers, LTE will put you onto IPv6 most of the time with short-term leases
> of IPv4 addresses when you need IPv4 connectivity. The network will remain dual-
> stack.
>
> So, no, you are not currently using some bizarre ipv6-ipv4 proxy back at the
> cell company NOC or anywhere else. At least not yet.
>

Sprint does their best to hide this from the general public but I can 
provide screen shots if needed.

>> Or as Homer Simpson would say,
>>
>> Mmmmmmm... bizarre hacks
>>
> lol
>
>> Seriously, it should be obvious that the economics of rolling out a brand new technology that is going to use IPv6-only plus a bizarre
>> hack to access the IPv4 Internet, is going to guarantee that the
>> bizarre hack is going to be hacked on until it works quite well.
>>
> Why? Why not instead work towards a much cleaner solution of eliminating the need
> to access the IPv4 internet? If the content and services people want are available on
> IPv4 and IPv6, then, there's no need for bizarre hacks to allow ipv6-only clients to
> reach IPv4-only content.
>

As long as there's some content on the Internet that's IPv4 only
the consumers of this technology are going to demand access to it.

>> NAT is a bizarre hack, wouldn't you say?  Yet most users are
>> happy with it.
>>
> Most users are happy with a great many things that are neither in their best
> interests nor necessarily good for the community. Most users are happy to
> keep putting gasoline in their automobiles, ignoring the damage being
> done as BP "brings oil to america's shores" as we speak. For a long time,
> most users were happy to try and get over viral diseases using antibiotics
> even though they had no positive effect against the virus and helped to
> breed antibiotic-resistant strains of bacteria in the process.
>
>> I think the issue here is not that bizarre hacks will create horribly
>> degraded IPv4 connectivity.  I think the issue is that bizarre V6-V4
>> hacks will get institutionalized, which will make it a lot more difficult to ultimately drop IPv4 and go IPv6 only.  That is a separate and valid concern, but FUDing it around isn't going to help anything.
>>
> I think that both are valid concerns, but, my more immediate concern
> is that bizarre hacks will create horribly degraded IPv4 connectivity with
> a second order effect that user perception of acceptable will move from the
> current moderately degraded situation to something even worse.
>

:-)

>> Technology companies have a long history of making bizarre hacks
>> work.  Just look at Microsoft Windows, one of the most bizarre hacks
>> in the history of technology (followed closely by Mac OS 6, 7, 8&  9.)
>>
> I think your statement here makes my point.
>

I don't see a problem with "new eyeball users" on IPv6 from accessing 
IPv6 content.  But I just cannot see how anyone deploying anything 
marketed at these "new eyeball users" can fail to ALSO supply a "bizarre 
hack" that works well and that allows the new eyeball users to ALSO 
access IPv4.

And as long as people are providing such bizarre hacks to the new
eyeball users, then people with IPv4-only networks will not feel
strongly compelled to switch

It's a classic Chinese finger puzzle.  People who are IPv6 need to get 
access to legacy IPv4 because not everyone is dual-stacked.  But, the 
major incentive to dual-stack is to service the new IPv6-only people.

There's really only 2 ways to break out of it.  Either you create a
significant group of IPv6-only users, or you create a significant
amount of valuable content only accessible on IPv6.  Either way would
do it, but the users want content that's on IPv4-only, and the
content providers want users who are on the IPv4-only network, so
both groups have a disincentive to switch to IPv6-only.

Ted

> Owen
>
>