IPv6 Readiness Report

MythoBeast writes "In the latest episode of the Intellectual Icebergs podcast, Brett Thorson of Ravenwing provides a very good review of how ready our industry is for IPv6. He also provides a pretty good implementation guide for those who want to set up IPv6 at home."

Spam must be controlled

[humankind]

We can't move to IPv6 until the spam problem is solved. With the additional address space that IPv6 offers, spam will increase by a googol if the spam gangs are not stopped. More spam is stopped because of RBLs now than any other method. IPv6 would make that obsolete.

IPv6 Design Mistakes

[Eravnrekaree]

I believe that the design of IPv6 was flawed in ways that it has inhibited adoption which could have been much more rapid. The IPv4 address space should have been a subset of the IPv6 address space. This would allow easy interconnectivity to Ipv4. The other direction, for going from Ipv4 to Ipv6 is trickier, but could involve manipulation of DNS. When a ipv4 peer requests a IP for a DNS address, the DNS server will reply with a private IPv4 address, the router/gateway associated with the DNS server will catch the connection to this IP and reroute the connection to the proper IPv6 address. It does only work with DNS addresses, yes. A special block of Ipv4 addresses should have been set aside for this purpose exclusively. Problem solved. Most people use DNS anyway. Other solutions could be devised to access a ipv6 address without DNS from ipv4, a protocol that would allow users to configure a forwarding route on the router via some utility, so that all connections to a private IP are rerouted to a specified IPv6 address. This could have eventually been built right into clients as well. This would have allowed a gradual switchover. The problem with the current switchover plan is that since there are so few Ipv6 users, there is not much incentive for websites to make themselves accessible on ipv6, but at the same time, users see no benefit from moving to ipv6, since there are not many websites avialable from it. So in order to access the internet, people need two seperate Ip configurations, people are not going to bother with ipv6 since it is pointless to them, all of the websites are on ipv4. Thus we get nowhere. It is absolutely true that there must be a gradual transition period where both protocols will be used and where both protocols must be interoperable.

Now I'm no CCIE but I like ipv6

[Anonymous Coward]

The more I get used to it and familiar with it. It's nice.

It's quite a bit more simple than IPv4. More importantly, last time I checked, the defaqult route tables were over 180,000, not just any old router can store all of that. Some of the plans for routing IPv6 based on geography will be nice and allow us to really scale the internet performance wise. No NAT, that alone dramatically simplifies a lot of things.

I think when Vista comes out the push will really begin. Comcast and other major ISPs are all readying their plans to roll it out. I for one welcome out 128bit overlords.

Re:Private networks and the business case.

[AgentGibbled]

Well, end-to-end connectivity would certainly make VOIP solutions considerably less hacky. Is that a bulletproof business case? Probably not, but it's an example of a useful application and it took me a couple of seconds to come up with it. I'm sure there are others if one were to actually think about it.

While I don't claim to be the world's leading expert on IPv6, I don't believe (and someone please correct me if I'm wrong) that it makes routers, proxies and firewalls go away. It does make NAT kind of redundant, but it doesn't seem to me as though that has much (any?) of a negative impact on security provided there is a proper firewall in place. It just means that the router doesn't need to do another lookup on each packet to figure out where it's actually supposed to go. NAT works as a stopgap measure, but it won't prevent the inevitable from eventually happening.

IPv6 Business Case

[netrangerrr]

There was no business case for the transition from ARPANET's old NCP protocol to TCP/IPv4 in the 1980s - but there were technically compelling reasons. Luckily the ARPANET pioneers realized that a new protocol was needed to easily integrate the new services and applications they were thinking of deploying. Soon the WWW, e-mail, etc. exploded as they were simple to deploy on a powerful TCP/IP infrastructure. IPv6 makes it cheaper to deploy new network services and applications (like imbedded IPsec and QOS routing) by adding new extension headers to define new services. It also scales massively and offers both private networks and E2E options. You'd be amazed at how much extra code/infrastructure is necessary to get around NAT today to make many applications work.

We are currently working on a paper, with help from subject matter experts of the North American IPv6 Task Force, on HOW to get a return on investment from IPv6 technologies by adding new IPv6 based network services to enhance reliability, security, QOS, and mobility support in networks.

Re:IPv6 isnt really wanted

[Anonymous Coward]

I think the big implication of IPv6 is in cheap wireless internet communication. Imagine a cellphone that's basically a VOIP phone, with IP access everywhere. There are hundreds and hundreds of millions of wireless phones around the world, such that it would break IPv4 with the way it's currently subdivided--what with some schools having more addresses than entire countries with populations rivaling that of the the most populus locations in the US.

Go to a third world country, and they might not have television (they often share televisions), inside toilets or even be able to afford food, but damned if nearly every home dosen't have one or two wireless phones... It's the same thing from the western tip of Africa to the furthest reaches of the Philipines; cellphones all over the goddamned place... Millions of 'em.

I don't know how it is in the rest of the world, but it seems that here in the US, pricing for data communication is, quite frankly, absurd... And what's the difference between voice and sendig an email? Voice data takes up thousands of times more bandwidth than sending an email, but it's priced at least a hundred thousand times lower, bit for bit! It's insanity, just like it was with ISPs before AOL practically forced everyone to go for unlimited minutes, for a reasonable monthly fee!

Going with IP for phones would open doors for all sorts of cool functionality. Wouldn't it be killer, for example, if your phone would allow use as a full time wireless router (bluetooth, or some other encrypted channel) for your computer or PDA, full time, and still be able to receive a call whist doing it? Voicemail could be revolutionized. It would make it affordable to use the functionality that's already built in to get stock quotes, browse the web, receive music from the provider's music store, or any myriad of things. Videomail and videophones could, for once, be possible, and accessible!

I know cell providers are scared of this, because it will relegate them to the status of electricity providers, metropolitan water, etc. They probably sit awake at night because this is their nightmare, and because it will put an end to the insane profits they experience... But the first one to offer IP cellphones at a reasonable price will slaughter the rest... I gurantee it. The only problem? IPv6 will be required, and with the built in QOS, encryption, etc, and it's the only thing that can get the technology moving world wide.

Today's Internet should be trivial.

[jd]

Most home users use DSL or cable modems and the ISPs would be quite capable of pushing new firmware to those to become IPv4/IPv6 gateways. You can then convert the entire "real" Internet to IPv6 without home users ever having to lift a finger.

Once that's been done, it's just a case of those same ISPs offering a CD to accelerate Internet usage (ie: which use native IPv6 rather than the gateway) and conversion is complete. Complete conversion of the Internet, by converting each ring in turn transparently to all outside layers, should be possible over the course of a few months at most. A solid concerted effort could probably achieve everything up to the end-user level in a matter of weeks, without a single person realizing what was happening.

Of course, I don't seriously expect that to happen. Not because it can't, but because the level of cooperation needed is likely beyond most businesses today. It's purely a political problem, not a technological one.

Industry is ready as "sort of"

[layer3switch]

Verizon DSL (NYC) not ready. Oh so NOT ready. CableVision (NYC) so not ready. All of my old linksys routers don't even support IPv6. Only thing I have ready for IPv6 is my damn Linux box.

Yeah, so far, I can ping myself all day... I'm just getting myself ready... any day now... really... c'mon... do it. do it.

Multihoming?

[Anonymous Coward]

Did they ever fix multihoming with IPv6? Last time I looked it was nearly impossible.

I suspect we will end up in a situation where the "rest of the world" uses v6 and the US uses v4 practically forever.

Re:Written guides for what?

[Anonymous Coward]

I think that tunnels are falling out of favor compared to 6to4 as a transitional strategy. With 6to4, everyone with a single IPv4 address gets a huge IPv6 address space to play with.

The use of the anycast 6to4 relay routers seems to be pretty trouble-free these days. I've set up 6to4 on hosts on Asian home ISPs and at US universities and gotten equal or better routing to the direct IPv4 route that is available (in terms of ping-determined path length and latency)!

There is the problem of some US ISPs and/or junky SOHO routers not being able to forward IPv6 traffic through their crappy NAT layer. Eventually, the only obstacle to IPv6 will be those providers who see it as a threat to their antiquated business models.

Re:IPv6 isnt really wanted

[big ben bullet]

i for one can't wait for the GUIDs to run out

i've got a dedicated dual core amd64 4000 with 4 gigabytes of ram at home constantly generating new GUIDs and storing them in an oracle database on a 10 gigabytes storrage array (expandable if necessary)

that way, when the world runs out of GUIDs i'll make a fortune selling them

otoh i hope the G stands for global and not for galaxy, or i could be in big trouble using up the GUIDs from the other side of the universe... i wouldn't want to provoke an intergalactic war because of entire nations running out of GUIDs or something

maybe i should start looking into this IPv6 thing too? afterall, if nobody really wants them, they're bound to be cheap for the time being

DJB Says...

[PhYrE2k2]

I'll just point everyone to DJB:
    http://cr.yp.to/djbdns/ipv6mess.html [cr.yp.to]

He pretty much covers most of it. IPv6 is dead on the public Internet long before it started. I knew this as soon as I called up MCI/WorldCom last year to ask if they had any IPv6 address space to add to our few class-C's and they laughed at me. If the folks who run half the Internet aren't ready for it, why would we be?

-M

Re:IPv6 isn't just addressing.

[mwood]

The people who don't care will be switched without knowing it, as soon as their suppliers decide that they want to or have to. If Microsoft decides that every XP user should have IPv6 enabled for some reason, the fix will come along through MS Update and you'll get it whether you know what it is or not. If your ISP decides that IPv6 is necessary, it'll be enabled whether your client requests an IPv6 address or not. When both have happened, hey presto! you have IPv6 and you didn't click a single button. Home-router manufacturers will lure most of their customers to swap out their old routers for new somehow...otherwise profits aren't sustainable...and IPv6 will come along for the ride when the vendor decides it's good for him.

"Consumers won't do it" is irrelevant. Consumers won't be asked. The few who never patch or upgrade will eventually find more and more applications dying or getting cranky, or they won't care because they never use new stuff and the old stuff still works okay.