As Christopher rightly points out IPv6 has built-in a lot of the features that have been "hacks" we have added on to IPv4
1. Lots of address space to waste - NAT not required.
2. Built in security (IPSEC) and Quality of Service at the IP level.
3. Mobility so you can connect via your "home" IP address.
The thing is, I think we have grown comfortable with the hacks and maybe the business drivers haven't materialised for it yet.
For instance, my story about an IP address in every light-bulb might seem far-fetched at the moment. But closer to reality might be an IP address for every mobile phone. It even makes sense - to be able to send and receive IP traffic to the same address no matter where am. The problem is that for mobile data services, you pay for traffic you both send and receive. And if your phone's IP address is just "out there" for some one to push traffic to you could be quite vulnerable. So the current method of "pulling" IP traffic from your, and hence just using a NATted (or at least non-associated address) makes more sense from a business point of view. So while from a utopian and ubiquitous connectivity point of view IPv6 mobile-IP makes perfect sense - unless bandwidth is free then it doesn't add much value on what we already have.
Also as far as government mandates are concerned, which Howard mentioned, unfortunately I have seen this fall in a big heap once before. There was a huge big, quite mature and technical quite good protocol stack called OSI (Open Systems Interconnect) that was developed by the commitees of the ITU. It was promoted very strongly by all and sundry and I think USA, UK and even Oz governments all launched programs called something like GOSIP (government OSI profile). These all had various timeframes for adoption. But unfortunately (for OSI) a grass-roots (and truly open) stack known as TCP/IP came to public knowledge in the early 90's and the rest (as they say) is history. I do come across OSI protocols now and then but you won't find many proponents of it. (Actually he biggest problem with OSI was it's cost. It was nearly impossible to find free protocol stacks. And this was because the standards and specification were all copyrighted by the ITU and had to be purchased. Therefore no self-respecting open-source hackers or academics were going to touch it. The IETF and the free nature of RFCs etc to my mind is prime driver for the wonderful spread of internet and it's releated technologies). I did a little googling for "ipv6 us government mandate" and there is stuff there - but not much. http://www.wcsr.com/default.asp?id=114&objId=181 reports that only slow progress is being made. Also the mandate at this stage seems to be only at the very core of the networks. My guess is that most will just be doing the bare IPv4 inside of IPv6 tunnelling that is required. This will be not much different from say what is happening with the core of ISPs transition from ATM to MPLS that has been happening over the last 5 years or so. This sort of thing has been totally transparent to end users (apart from some marketing from the carriers on their "next generation MPLS networks".)
But despite my nay-saying I still think we should be encouraging customers and suppliers to examine the future and at least make it available for our evaluation, as Christopher has suggested. It probably is about time I rattled by ISPs chain so I might do that. (Actually an area that I think you should really hassle them about is proper IP multicast and end-to-end quality of service. That way things like IP streaming of multimedia would actually work properly, and without having to either duplicate the same streams a gazillion times. The only problem is that it probably requires your ISP to talk to their upstream provider into providing the same service (and so on). This is probably a difficult thing)
Enough musings for now, regards Martin
-- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
