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."

Comments

IPv6 isnt really wanted

[Anonymous Coward]

IPv6 is a solution looking for a problem, at the moment in its current state nobody will use it, its complex , doesnt play with legacy systems (even win2k support is flaky at best) all those routers and wifi boxes that best buy are selling, most of the ISP's dont want it and dont support it let alone the users figure it out

its another "its coming" technologies thats "nearly" with us for the last 10 years and STLL nobody really cares, its like W3C validation, nice in theory but most people dont care about it and most of the html generation tools dont create it

Re:IPv6 isnt really wanted

[shawn(at)fsu]

Just wondering is it better to fix a problem before it arises or wait until it's about to bite you. I'm thinking of the /. issue with VIN's to run out soon It wasn't really a failing of VIN as it achived what it's goals were for the required time. Can't some of the same be said about IPv6.

Re:IPv6 isnt really wanted

[cgranade]

It is wanted, as it solves a very pressing issue. With more and more mobile devices and embedded devices requiring their own IP addresses, we are running out of address space. Furthermore, the design of IPv4 relies upon assumptions that are no longer valid, nessesitating such ad hoc and stop gap solutions as NAT. While NAT may be useful in its own right, it should not be used solely to allow for more devices.

As for the comment about W3C validation, it always has been, continues to be and will most likely continue to be very important in the future. Without such a service, how is one to tell what XHTML, HTML, etc. actually are? Machines are not intelligent, and so we cannot be content with the tag soup that passes for HTML on most sites, but we must reqire some sort of standard for quality. I would love to see a browser that, by design, will choke on any non-validating input, since by design such a browser would be simpler and easier to maintain. Without quality control mechanisms such as W3C validation, we would have a very poor Internet indeed.

Re:IPv6 isnt really wanted

[Anonymous Coward]

> While NAT may be useful in its own right, it should not be used solely to allow for more devices.

Umm, that's precisely why it's used. So it doesn't adhere to the purity of the end-to-end argument (in fact, it pretty much smashes it), big deal. It works, and it's the defacto standard, and it's pretty much pushed off the need for IPv6 to the unforseeable future.

Re:IPv6 isnt really wanted

[hhr]

IPV6 suffers from the another-technology-is-good-enough-and-cheaper problem.

Beta was superior, VHS was good enough and cheaper.

Audiofile stereo equipment is superior. An IPod is good enough and cheaper.

IPV6 is superior. IPV4+NAT is good enough and cheaper. Which is very unfortunate because IPV6 solves real problems.

Re:IPv6 isnt really wanted

[DDLKermit007]

Actually NAT serves us quite well in our situation. Cellular devices (mainly from China) are the big pressing fricking issue here and for the most part cell phones do NOT need real public IP space. There are extremely far and few betweens where a cell phone from any nation needs an IP that can be pinged from the outside or otherwise accessed. Cellphones make thier own calls out to the internet and negotiate a way for the data to be sent to them. Only in the case of network present apps and say Crackberries does a private IP space make allot of sense (of which can be worked around eaisily).

IPv6 is too big & complicated and does not play well with older systems (another poster noted Win 2K support is flakey at best). Do you honestly expect older devices like cellphones to be updated by the manufacturers or even better those of us using Treo like devices where we don't just throw them away each year and get a new one. IPv6 would "work," but it's not the thing thats going to work "best" (for one good luck keeping a list of 50+ IPv6 IPs memorized).

As for W3C quality control is involved I and many others would love that kind of setup. However that would block off many people who fit into the "I can code 1337 HTML for my grandma" family, but not the "I can learn to code well" group. Hell the internet hit critical mass because of browser & network flexability and not ridgidness and "quality control." Not everyone can code HTML as well as "some" on /.

What'll likely happen is all cellphones will migrate twards IPv6 (or something like it that works better) with a NAT between all of them and the rest of the IPv4 network and as older devices running the old IPv4 stack get older and older (old cells, 95, 98, ME, 2K, old Mac OSs) we'll slowly get over to whatever new thing. IPv6 is like HD-DVD & Blue Ray. Sure they might be nice as they are for the most part they are too soon and not just right, but at least IPv6 doesn't require hardware & licencing deals that can bankrupt companies when it false starts.

Re:IPv6 isnt really wanted

[jamesh]

I'm right now struggling with the various implementations of NAT-T (IPSEC NAT Traversal) and the fact that they won't play nice together. Wouldn't be necessary with IPv6.

Ever tried to set up a VPN between two sites which both use 10.0.0.0/24 as their network range?

Ever wished you could just ssh direct to your desktop machine from home without futzing around with vpns?

So you may not want it or see the need for it, but if you understood the amount of work that has gone into making NAT the 'solution' it is today you might appreciate it a little more :p

Re:IPv6 isnt really wanted

[toddbu]

I'm just not sure that's true. It certainly seemed that way when IPv6 was invented, but since then NAT has become a regular feature on home and business networks. Add in the regular use of DHCP to autoconfigure devices to a network, and you find that there's no longer any real pressure to make the switch to IPv6. Thus it made a lot of sense when it was developed, but now it seems pointless.

It may be pointless to you, but there are many people who could deparately use it. Think of all the problems that go away when NAT is gone. Like being able to use BitTorrent or SIP or any other "push" technology without having to set up port forwarding on your router. And even when you do get it set up right, you can't run on multiple machines behind a firewall without some kind of proxy on the other side. NAT is to the Internet was segmented memory was to CPUs - a great idea to move things forward but not a good long term solution.

I'm really jazzed about the idea of having my own personal 64 bit address space on the Internet. Then again, I'm not sure that even that will be enough. :-)

Re:IPv6 isnt really wanted

[bigpat]

big deal. It works

Ummm, no it doesn't work. It works for a few things, and breaks a whole lot of other things. You are arbitrarily limiting a whole set of end-to-end applications simply because you have no imagination. The simple fact is that I can, with my static IP, do a hell of a lot more than you can with some short leased DHCP IP behind a NAT.

Re:IPv6 isnt really wanted

[frakir]

there is no legitimate reason for wanting to access a mobile phone remotely.
hmmm............

Re:IPv6 isnt really wanted

[ultranova]

The only devices that need public IPs are servers. Hell, it's a potential security hole to give a non-server a public IP *at all*.

In Internet, every device is a server. That some of them are dedicated specifially to server duties does not change this. Filesharing networks, netphones, anything that lets two machines to exchange information in realtime - they all require at least one machine to have a public IP so it can be contacted. So yes, in Internet, every device needs public IP in order for the network to function.

Of course there are many interests that would love to see Internet to get broken and replaced by old-style broadcast network, since that would stop the competition from independent parties to those interests power. RIAA and MPAA, as two best examples, want to close Internet as a distribution channel for anyone but themselves. ISPs don't want you to be able to run your own servers, since that will increase the bandwith consumption and therefore decrease their profits. Blizzard and other MMORPG makers want to keep the costs of running a (small) server ridiculously high to keep competition to a minimum.

These are the real reasons for dynamic IPs, port blocking, and NAT. They are inconvenient, because they are designed to inconvenience you, to keep you in your role as a consumer. Producers don't want competition, and will do anything to stop it from happening.

Mobile phones for example do *not* have public IPs and never should do - there is no legitimate reason for wanting to access a mobile phone remotely.

Unless, of course, you want to call one ;). IP address is simply the Internets equivalent to a phone number.

Like Y2K?

[microarray]

Could someone tell this uninformed person what the hype is all about? So, we run out of IP addresses, so what? Seems like a market then exists where you could on-sell your IP addresses for $$$. Prices go up too high, market forces then result in IPv6 implementation. What's the problem?

Re:A podcast guide?

[daniel23]

I agree with this, unlike a written guide a podcast has no copy'n'paste and it is much harder to follow talk than written text when the language used is not your native tongue.

NAT provides a firewall

[tepples]

Umm, [adding more devices is] precisely why [NAT is] used.

Apart from that, NAT is also useful because of an inherent side effect, namely that a basic firewall comes "free" once your router has implemented NAT.

Private networks and the business case.

[zerofoo]

It has been said many times here on Slashdot, but it bears repeating.

There is no business case (yet) for IPv6. The internet was designed for resilient point to point connectivity, but the business world does not want that.

Today's security paranoid businesses want to keep their internet exposure to a minimum. Look at most companies - lots of computers behind one or two public IP addresses. Most internal hosts are firewalled, proxied, and natted INTENTIONALLY.

Sure, this creates some problems, but there are workarounds for most issues.

I keep hearing about handhelds and that millions of them will need their own IP addresses. I don't see why. I'm sure most of the wireless providers want to control the content that their subscribers can send or receive - that business model does not want a wide open network with each host directly connected to the internet.

In this type of business environment, I can't see why any business would want to throw away thousands if not millions of dollars in their existing IPv4 investment.

If you can explain a bulletproof business case for IPv6, then Mr. Chambers at Cisco may have a nice sales job for you.

-ted

Re:Private networks and the business case.

[roystgnr]

I keep hearing about handhelds and that millions of them will need their own IP addresses. I don't see why. I'm sure most of the wireless providers want to control the content that their subscribers can send or receive - that business model does not want a wide open network with each host directly connected to the internet.

Back when it was just a proprietary BBS, Prodigy wanted to charge me $0.25 per email I sent - that business model does not want a wide open network where any host can connect to any SMTP server.

I think they became a full TCP/IP provider eventually, but I switched networks too quickly to find out. Let's hope that wireless providers understand the lesson here: if someone else can offer your customers a better business model, it doesn't matter what your business model wants.

Re:Private networks and the business case.

[AK Marc]

Most internal hosts are firewalled, proxied, and natted INTENTIONALLY.

Most internal hosts are natted. I'm not sure about firewalled, and certainly not that many proxied. NAT is not a security measure. It does provide some security, but then so does having oil poured all over your front porch - but neither was created for or ideal for security. NAT was created to connect previously unconnected networks. It was not created for security. Security is an idea that was tacked on to it later to explain why home users should have NAT. Security is a marketing feature for helping sell NAT devices. Using it for depleted address space came years after NAT was first used as well.

No, it was the networks in the beginning that weren't on the Interent that were properly addressed. They were given 192.168.0.0 172.16.0.0 and 10.0.0.0 addresses (so used because they weren't on the Internet, so the companies didn't need to pay for the networks to get them on). Years later, when there was a need for these privately addressed computers to reach the Internet, it was cheaper to use NAT than change the IP on all those computers. NAT was a bean-counter's solution for poor planning (or saving some money in paying for addresses that weren't going to be used on the Internet). NAT had nothing to do with depletion of addresses, nor security. But most forget about that now, since that is not a though in any of the use of it now, aside from the few times a company merges with another company with the same internal IP range and double NAT gets some action. That's closer to the initial intention than any other use I've seen in a while.

Re:By the time IPv6 is ready

[VGPowerlord]

One company does not an industry make.

Re:Business case for IPv6

[Anonymous Coward]

I don't get this argument at all. Since you mention connections and most interesting traffic is over TCP, what do fragments have to do with it. Someone could send TCP traffic one byte per IP frame. If the firewall is going to do validation, it will have to reassemble enough of the protocol stream to understand what it is seeing.

Re:Business case for IPv6

[jd]

The argument is that IPv4 is excessively complex. The header has a vast amount of information, much of which any stateful device will nee to check and validate. With IPv6, the extra information either doesn't apply (as in the case of fragmentation) or is pushed into secondary headers and only examined by layers that actually NEED to care.

Since I used fragmentation as an example, when is fragmentation important? Well, let's say Business A uses standard ethernet frames (1500 bytes) and Business B uses jumbo frames (6000 bytes). Business B's packets will be fragmented into 4 parts at the point where jumbo frames are no longer supported. They will be re-assembled into a jumbo frame on Business A's firewall (in order for the packet to be validated) and will then be broken up again as Business A's network won't support jumbo packets.

All that takes time. If a fragment is dropped, in transit, the jumbo packet won't reassemble correctly and will be dropped, forcing the entire jumbo packet to be resent. (In other words, a dropped packet is 4 times as expensive.)

With IPv6, that doesn't happen. Business B connects to Business A. Negotiation identifies that the largest packet that will travel intact is 1500 bytes, so Business B (when sending to Business A) will use packets of that size. No fragmentation, a drop will cost 1500 bytes not 6000 bytes, and it doesn't involve Business B reducing its MTU to anyone else, so if other people can receive jumbo packets fine, the connection isn't degraded.

It doesn't help that IPv4 is based around byte-alignment and bit flags, whereas modern computers assume 32-bit or 64-bit words. Having things word-aligned and word-sized is much more efficient on a modern computer. That is something that has genuinely changed over time and wasn't merely a case of really bad design.

Re:Why do we need to be ready?

[Jugalator]

IPv6 is not needed, NAT works.

For a fraction of what you can do on the Internet, yes. Stop oversimplifying.
Even I as a regular user have run into the problems with two NAT'ed people trying to communicate with each other.

IPv6 supports fragmentation

[Helevius]

That's hardly a "business case." And as another poster (unfortunately not being modded up) pointed out, IPv6 supports fragmentation. It's just that end hosts have to fragment and reassemble, and not intermediary routers. So, your firewall will see fragments anyway.