Re: RE: New Version Notification for draft-jiang-v6ops-incremental-cgn

[JiangSheng 66104 <shengjiang@huawei.com>]

> end users view networks as a business overhead expense and are 
> unlikely to needlessly spend money on networking technology 
> without a solid business motivation to do so. I went on to explain 
> that IPng is unlikely to be deployed by us end users unless it 
> becomes bundled with a business requirement, the most compelling 
> of which would be a new Killer Application that demanded IPv6 
> capabilities in order to function.

Fully agreed. This is particularly true for IPv6 deployment in last ten year - failure to find a killer application. A right business model is actually more important than technology itself.

> At this current time, IPv6 is very immature and IPv6 deployments 
> have very high risk when compared to IPv4 for the end user. There 
> are only negative business reasons for deploying IPv6 at this time 
> (i.e., I can articulate many compelling business reasons to NOT 
> deploy IPv6 but the only reason to deploy it in the USA today is 
> government decree -- which didn't work for OSI and is unlikely to 
> work alone by itself for IPv6.). There are currently no technical 
> reasons for the end user to prefer or want IPv6 over IPv4. 

I believe this situation is changed. IPv4 has reached its limitation. End-user will soon find that they could NOT obtain a public IPv4 address and with private IPv4 address (maybe more than one NAT in the way) they were only able to communicate with part of Internet world. ISP has met the problem that they cannot adopt new services due to the lack of public address.
 
> Unless this changes in the future, IPv6 will continue to not be 
> deployed by end users despite the efforts of Apple and Microsoft 
> and others to ease its adoption. But then, since not all ISPs 
> support IPv6 today and the DFZ Internet is currently unprepared 
> for the BGP scaling problems that would arise with a mixed IPv4-
> IPv6 infrastructure this is probably A Good Thing.

This is a real issue which has been ignored by IPv6 community for years. My guess is IPv6 community does not want to discuss this because it may even more block IPv6 deployment. However, it is time for us to face it and solve it or avaoid it.
 
> I personally *HOPE* that IPv6 will become deployed -- after all, I 
> spent many years helping to create it. But my experience with pre-
> TCP/IP protocols (I used to be an SNA and BSC "expert") is that 
> even if IPv6 becomes widely deployed, IPv4 will endure for a 
> surprisingly long time (decades). 

Yes, it will be years. I believe the whole internet society has accepted this and be prepared for a 25+ year co-existing period.

Best regards,

Sheng
 
> I have many conflicting visions of what this could mean. Back in 
> 1999 Tony Li privately explained to me how IPv4 could endure 
> forever with no IPv6 ever being deployed. Tony's insights are 
> commonplace today as evidenced by the combination of NATs, private 
> addresses, and VPNs. I admit that as a possibility but I hope that 
> it doesn't happen. I prefer for IPv6 to gradually become deployed 
> and replace IPv4. However, my clear preference is for the 
> flexibility of the RANGER solution. The RANGER solution creates an 
> interworking environment so that it interworking occurs no matter 
> what the transport is at any part of the Internet -- the network 
> can cleanly support IPv4 only, IPv6 only, or IPv4-IPv6 all mixed 
> together. Such flexibility is a powerful tool for handling 
> whatever vision we may have for the future of networking.
> 
> --Eric Fleischman
> 
> -----Original Message-----
> From: Gert Doering [mailto:gert@space.net] 
> Sent: Tuesday, May 19, 2009 12:25 PM
> To: Templin, Fred L
> Cc: Rémi Després; Sheng Jiang; Brian E Carpenter; 
> v6ops@ops.ietf.org; guoseu@huawei.com; Fleischman, Eric; Russert, 
> Steven W
> Subject: Re: New Version Notification for draft-jiang-v6ops-
> incremental-cgn
> 
> Hi,
> 
> On Tue, May 19, 2009 at 07:29:59AM -0700, Templin, Fred L wrote:
> > > I believe it should be quickly agreed on, to expand 
> applicability of 
> > > a useful tool for rapid deployment of native IPv6.
> > 
> > Native IPv6 is not necessary with a solid VET deployment.
> 
> I can't follow that line of reasoning.  The end result has to be 
> an IPv6-only network, not something held together with IPv4 clutches.
> 
> Native IPv6 without IPv4 is the only sustainable solution in the 
> long run.
> 
> Gert Doering
>        -- NetMaster
> --
> Total number of prefixes smaller than registry allocations:  128645
> 
> SpaceNet AG                        Vorstand: Sebastian v. Bomhard
> Joseph-Dollinger-Bogen 14          Aufsichtsratsvors.: A. Grundner-
> CulemannD-80807 Muenchen                   HRB: 136055 (AG Muenchen)
> Tel: +49 (89) 32356-444            USt-IdNr.: DE813185279
>