[arin-ppml] Legacy Space authority

[John Paul Morrison]

michael.dillon at bt.com wrote:
>> Maybe I'm reading too much into your statement, but you're 
>> implicitly saying that IPv4 addresses will have NO value in two years.
>>     
>
> They have no intrinsic value today. Things will get tight for a while,
> but once it is clear that IPv6 is relatively easy to deploy with
> reasonable
> capex/opex, I don't see how IPv4 addresses can retain any value that
> they *MIGHT* have gathered in the next couple of years. Some people
> are really pushing for IP addresses to be monetized, and it may
> happen to some extent, but if it does, it will collapse in about two
> years when the IPv6 technical issues are mostly solved.
>   
Let's not get into existential hair-splitting about value. Common sense 
says that IPv4 addresses have do have value in a number of ways.
They can be used for commercial purposes, and possessing the right to 
use/hold them gives the bearer ("owner" although many don't like the 
word)  some use and value, and possible resale value.

You might as well say that a $100 bill (or Euro) has no intrinsic value 
- well "obviously" they're just pieces of paper with no "intrinsic" 
value, but I can still do a lot more with a $100 bill than a blank sheet 
of paper.

Your original point was that a person would be foolish to put any value 
on IPv4 addresses in two years. Now you're saying they might at least be 
valuable as a hedge...

It isn't obvious yet that Telcos won't find a way to profit from the 
IPv4 exhaustion simply by extortion. Pay us more, or we can't guarantee 
IPv4 connectivity, or pay us more for IPv6 connectivity, but be able to 
talk to almost no-one at first. What would most people opt for? Anyone 
who gets their IP address today by DHCP will probably see their routable 
IPv4 addresses re-claimed or have to pay more for them. Will the 
broadband companies hold the hands of millions of users to upgrade to 
IPv6, or just proxy/nat? Will hosting providers pay to upgrade and jump 
through paperwork hoops when those same broadband companies still keep 
bringing them to their door?

>> I'm sorry if it sounds 
>> cynical, but I just don't see vendors, ISPs, or businesses 
>> demanding that they or anybody else ready for IPv6 *now* for 
>> what might happen in two years.
>>     
>
> Rip van Winkle, eh? Have a look at Cisco's and Juniper's website
> and you will see that they have been doing IPv6 ready for years.
> Look on ARIN's wiki to find the April 2nd article in Network
> World that tells what several of the largest US ISPs are 
> doing. http://www.getipv6.info/index.php/IPv6_in_the_News
>
>   
Hardly. I've been following IPv6 since TUBA, and remember listening to 
Steeve Deering in the mid 90's presenting the just finished IPv6, and 
saying it would take a few years. I don't recall the time frame, but I 
recall him mentioning that some network had just finished ripping out 
XNS (or whatever) after twenty years, and that IPv4 would be with us for 
a time.

I've read the IPv6 docs, as well as implemented it for testing,  and am 
very aware of what the vendors are doing. There's also lots of docs on 
how to setup CLNS and IS-IS, which was mandated by Government 
procurement regulations. I believe the US GSA has mandated IPv6 as well, 
so no surprise that there's IPv6 support.
Vendors are simply marketing the features, you won't find vendor sales 
or consulting engineers pushing IPv6 up front or telling Enterprises 
that the wolf is at the door. IPv6 is a footnote, nice to have.

I'm not knocking IPv6 - "I want to believe"!  But I just don't see IPv4 
collapsing, I just see the ISPs cannibalizing their networks and 
kludging things up to keep business working as usual.

IPv6 could probably have been (and maybe should be) just a seamless, one 
line feature on your boxes, and at one point renumbering and other 
transition mechanisms were proposed but never materialized. There's 
short term tunnels, fixes and workarounds but not really an "easy 
button" yet, at least for the end user.  I would be more hopeful 
predicting that vendors, OS makers and customers (aka the market) will 
form a loose consensus on a seamless way to turn on IPv6, than I would 
be in predicting the IPv4 Internet is going to end in two years, which 
seems the only way IPv4 addresses will cease to have value.


> Same website has this page
> http://www.getipv6.info/index.php/IPv6_at_ARIN_XXI
> with links to various work being done right now to iron out the
> interworking
> issues between IPv4 and IPv6.
>
> Before you know it some market analyst will start asking telecoms CEO's
> what they are doing to prepare for the IPv4 runout in 2010 and share
> values will dive (or not) based on the answers. 
>
>   
Important but probably academic. Business will not cease for those 
telecoms because IPv4 is not going to drop dead, they can shift address 
space around - and make extra money on top of it all, and because most 
can reasonably say to auditors that they could implement IPv6 if they 
had too.
>> You'll probably just see proxies/caches/NATs being used to 
>> expand IPv4 address space, and anyone launching a new web 
>> service is going to need to find someone to host it on an 
>> IPv4 server, or their business is going to go bust, because 
>> they will be talking to a small crowd of IPv6 users, or only 
>> aiming at users in emerging markets/developing countries.
>>     
> Hosting providers can run 6to4 gateways so that v6 websites
> are fully accessible. If you are going to invest in proxies
> caches and NAT, then NAT-PT seems like the most risk-free way
> to do that.
Of course - but will hosting providers be the ones spending a lot of 
money for appliances and licenses, not to mention upgrades/changes to 
routers and Internet feeds, to handle a trickle of new IPv6 users, or 
will it be business as usual for some time? Or will it be the DSL and 
Cable ISPs with huge existing IPv4 allocations, who already have 
proxies?  Won't they be the ones with an economic incentive to conserve 
IPv4 addresses and re-market or "sell" them?  Will a routable IPv4 
address suddenly become a "business package" service for DSL/Cable 
users, at a premium price?

Whatever happened to all those IPv6 addresses on millions of cell phones 
that was supposed to drive IPv6? Seems like proxying works fine and 
profitably for the mobile telcos.



 



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20080505/7e65f046/attachment.htm>