Re:A crisis in the making: Only 4% of the Internet supports ...
by dfranke on Friday March 06, 2009 @02:05PM
(#505815)
Fears about IP address exhaustion are overblown. X-day is the day that IP address scarcity begins to become a problem, not the day that the intertubes burst and flood everyone's basement. We appear to be headed for an IP shortage because IP addresses are presently distributed by a central authority, "to each according to his need". IANA allocation policies undercut the fair market value of IP addresses, so the free market that could exist for them currently doesn't. But once IANA stops handing them out, there's nothing keeping it from forming.
Re:A crisis in the making: Only 4% of the Internet supports ...
by jeremyw on Friday March 06, 2009 @01:42PM
(#505784)
Vint Cerf and others intimate full interoperability between IPv4 and IPv6 (i.e. I can get to every site, regardless of my network) was beaten to death in the standards committees, but ultimately there was no technical consensus to be had. Unfortunately, it was a mandatory component to adoption. How did they convince themselves it would happen otherwise? They failed.
Now we'll have a clusterfuck of overlay technologies, dual-stack corner cases and last-minute crises. Yay, IETF.
Re:A crisis in the making: Only 4% of the Internet supports ...
by hapless on Friday March 06, 2009 @03:19PM
(#505900)
This guy looks at 4% uptake and says the sky is falling. I look at 4% uptake and say we're in really great shape.
According to this guy's somewhat pessimistic view, we're at least two years from the beginning of IPv4 shortages. We already have 4% uptake when there's very little incentive to do anything about ipv6. When the pressure comes on a little bit, uptake will skyrocket.
The great thing is that there's a network effect here: the more people get on ipv6, the more benefit there is to each additional punter. It's more and more attractive as more people join. I think we'll have no problem rolling this snowball.
Re:A crisis in the making: Only 4% of the Internet supports ...
by pmjordan on Friday March 06, 2009 @11:36AM
(#505612)
I keep meaning to set my local network up for IPv6 and then asking my ISP if they can hook me up with an IPv6 range. And if not, keep pushing. I think that will have to be the way forward. Application support is actually pretty decent at this stage, it's infrastructure that's the main problem. Oh, and I'll probably have to upgrade my DSL modem...
Re:A crisis in the making: Only 4% of the Internet supports ...
by tptacek on Friday March 06, 2009 @11:46AM
(#505623)
Application support in the major mainstream software market isn't horrible, but support outside of it is awful. The cost of porting software to IPv6 is a major hidden drag on adoption. Which doesn't bother me much, since I'm not sold on IPv6 at all.
Re:A crisis in the making: Only 4% of the Internet supports ...
by pmjordan on Friday March 06, 2009 @12:40PM
(#505697)
Out of interest, what alternative solution would you prefer? Starting to use NAT and private/unrouted address spaces for everything seems like a terrible way to deal with the situation. (and I've had various problems with VPNs as a result of it; learnt my lesson and use "rarer" subnets like 172.16-31 and 192.168.x where x != {0, 1, 2, 10, 100} now) I'm surprised nobody seems to be suggesting we break up some of the Class A networks - IBM, HP etc. surely don't need 16-32M addresses.
I admittedly know very little about the shortcomings of IPv6; I do know it handles fragmentation differently, which is apparently causing problems with tunnels, and as far as I can tell, it's a much "bigger" standard, not just IPv4 with a bigger address space, with some parts that are possibly completely pointless. (a little like multicast in IPv4 which seems to get very little practical use) As a developer, is there any particular technology other than IPv4 and IPv6 that I should be targetting?
Re:A crisis in the making: Only 4% of the Internet supports ...
by tptacek on Friday March 06, 2009 @03:05PM
(#505883)
I think that one of the major architectural failings of the IETF over the past N years is the belief that we should be building on and extending the core Internet protocols, when what we should be doing is moving more and more logic into the application layer.
So for instance, if you want to be able to address billions of people, the scheme that you use to do that should look more like email addresses (which are effectively unbounded) and less like a globally synchronized database of integers.
A lot of problems that are basically intractable at the router/switch layer are much easier to solve with relay servers. Multicast is another great example; IP multicast was a failure, but multicast delivery semantics with (say) 80% of the performance improvements of IP multicast are achievable with relays.
This isn't a crackpot idea; Frans Kaashoek and rtm at MIT had a project investigating it, called Reliable Overlay Networks. I did a startup in the late '90s based on overlays. The technology worked great.
The pain involved in NAT and 1918 addresses would be mitigated if our applications were less tightly coupled to IP (gethostbyname -> sockaddr_in -> connect), and if we could just pick a NAT traversal / NAT rendesvouz technology and run with it.
So my answer to your question is, you shouldn't worry about the alternative. Just keep building normal socket applications.
Re:A crisis in the making: Only 4% of the Internet supports ...
by jrockway on Friday March 06, 2009 @01:14PM
(#505743)
I get the impression that tptacek hates everything that wasn't either invented in the 1950s, or that he didn't invent himself.
IPv6 is the future of IP, regardless of its flaws. It is much farther along than anything else, and I don't see that changing. (And I don't see anyone using IPv6 until IPv4 addresses become Really Expensive, which probably won't be for 10 or 15 more years.)
Re:A crisis in the making: Only 4% of the Internet supports ...
by electromagnetic on Friday March 06, 2009 @04:23PM
(#505968)
I don't see the urgency being a matter of years, due to the fact that most ISPs are adopting NAT's. I guess my problem is, I don't see why ISPs won't just start deploying NAT's to convert IPv4 to IPv6. I mean with the span of IPv6 addresses, why can't they pin an IPv6 to an IPv4?
Re:A crisis in the making: Only 4% of the Internet supports ...
by shimon on Friday March 06, 2009 @01:38PM
(#505778)
Exactly right. It's not painful enough to bother yet. When it does get painful enough, maybe IPv6 will still be the frontrunner, or maybe something else will emerge, but we should not expect to figure this out for a few years.
Re:A crisis in the making: Only 4% of the Internet supports ...
by krschultz on Friday March 06, 2009 @12:00PM
(#505647)
I've seen and agreed with this warning several times, but today I realized I have yet to do anything about it. My router and my machines are all using IPv4 and I haven't a clue as to how well my ISP supports it. What about everyone else here? If were not on it in any numbers yet, the general population will take many years.
Re:A crisis in the making: Only 4% of the Internet supports ...
by electromagnetic on Friday March 06, 2009 @04:09PM
(#505952)
I've long doubted this as a serious problem. With the amount of packet filtering and what not that ISP's can manage on the fly, why can't they simply use a NAT to translate an IPv4 into an IPv6? It would fix the problem with most NAT's as both IPv4 and IPv6 can be tied together allowing a permanent IPv4 address to be a IPv6 on the net.
Re:A crisis in the making: Only 4% of the Internet supports ...
by metatronscube on Friday March 06, 2009 @04:10PM
(#505953)
No, this is not a problem (speaking as a Network Design Engineer) these kind of articles are completely overblown and based on nothing but fantasies and hype. Address exhaustion can be very easily controlled at the ISP-NAT level and we are already building in IPv6 capable devices in our networks...its not a problem.
Re:A crisis in the making: Only 4% of the Internet supports ...
by igglepiggle on Friday March 06, 2009 @12:03PM
(#505651)
Maybe the push towards DNSSEC will sweep IPv6 along with it.. we can't afford not to have either of these things, and both require some significant maintenance work - why not roll out both at the same time? Maybe DNSSEC should only support IPv6?
Re:A crisis in the making: Only 4% of the Internet supports ...
by tptacek on Friday March 06, 2009 @03:08PM
(#505886)
Couple things.
First, the "push towards DNSSEC" is largely just a superficial trend story. There are operational problems with DNSSEC that will probably severely retard DNSSEC adoption. I'm not going to pretend DNSSEC isn't more likely to happen after the last few months, but none of the problem with DNSSEC that almost left it DOA before have changed.
Second, there's no real relationship between DNSSEC and IPv6. DNSSEC signs arbitrary records, including A.
Re:A crisis in the making: Only 4% of the Internet supports ...
by tptacek on Saturday March 07, 2009 @11:43AM
(#506552)
Again, a couple things. First, the CircleID site is a proponent of DNSSEC. Second, this isn't the first time Verisign has publicly committed to DNSSEC. Finally, just because Verisign says it's going to adopt the protocol, doesn't mean the protocol is going to work.
Now we'll have a clusterfuck of overlay technologies, dual-stack corner cases and last-minute crises. Yay, IETF.
How well Dan's article has stood up: http://cr.yp.to/djbdns/ipv6mess.html
According to this guy's somewhat pessimistic view, we're at least two years from the beginning of IPv4 shortages. We already have 4% uptake when there's very little incentive to do anything about ipv6. When the pressure comes on a little bit, uptake will skyrocket.
The great thing is that there's a network effect here: the more people get on ipv6, the more benefit there is to each additional punter. It's more and more attractive as more people join. I think we'll have no problem rolling this snowball.
I admittedly know very little about the shortcomings of IPv6; I do know it handles fragmentation differently, which is apparently causing problems with tunnels, and as far as I can tell, it's a much "bigger" standard, not just IPv4 with a bigger address space, with some parts that are possibly completely pointless. (a little like multicast in IPv4 which seems to get very little practical use) As a developer, is there any particular technology other than IPv4 and IPv6 that I should be targetting?
So for instance, if you want to be able to address billions of people, the scheme that you use to do that should look more like email addresses (which are effectively unbounded) and less like a globally synchronized database of integers.
A lot of problems that are basically intractable at the router/switch layer are much easier to solve with relay servers. Multicast is another great example; IP multicast was a failure, but multicast delivery semantics with (say) 80% of the performance improvements of IP multicast are achievable with relays.
This isn't a crackpot idea; Frans Kaashoek and rtm at MIT had a project investigating it, called Reliable Overlay Networks. I did a startup in the late '90s based on overlays. The technology worked great.
The pain involved in NAT and 1918 addresses would be mitigated if our applications were less tightly coupled to IP (gethostbyname -> sockaddr_in -> connect), and if we could just pick a NAT traversal / NAT rendesvouz technology and run with it.
So my answer to your question is, you shouldn't worry about the alternative. Just keep building normal socket applications.
IPv6 is the future of IP, regardless of its flaws. It is much farther along than anything else, and I don't see that changing. (And I don't see anyone using IPv6 until IPv4 addresses become Really Expensive, which probably won't be for 10 or 15 more years.)
First, the "push towards DNSSEC" is largely just a superficial trend story. There are operational problems with DNSSEC that will probably severely retard DNSSEC adoption. I'm not going to pretend DNSSEC isn't more likely to happen after the last few months, but none of the problem with DNSSEC that almost left it DOA before have changed.
Second, there's no real relationship between DNSSEC and IPv6. DNSSEC signs arbitrary records, including A.