No IPv6 Doomsday In 2012

itwbennett writes "Yes, IPv4 addresses are running out, but a Y2K-style disaster/frenzy won't be coming in 2012. Instead, businesses are likely to spend the coming year preparing to upgrade to IPv6, experts say. Of course there's a chance that panic will ensue when Europe's RIPE hands out its last IPv4 addresses this summer, but 'most [businesses] understand that they can live without having to make any major investments immediately,' said IDC analyst Nav Chander. Plus, it won't be until 2013 that North America will run out of IPv4 addresses and there's no sense getting worked up before then."

Comments

Business as usual

[InterestingFella]

ISP's and hosting companies will not run out of IPs. This only means that the price per IP will start to slowly grow. Hell, every time I order server the companies still happily hand me over 5 IPs without me even asking for them. With a simple request I can also buy 256 ips for the price of $300 a year.

Re:Business as usual

[Anonymous Coward]

a Y2K-style disaster/frenzy won't be coming in 2012. Instead, businesses are likely to spend the coming year preparing to upgrade to IPv6

Sounds just like a Y2K-style disaster/frenzy.

Re:Business as usual

[Shakrai]

why do you even need a static IP for your home?

The question I'd like to ask is why don't they just hand out a static IP by default? The vast majority of broadband connections are always on; they aren't saving IPs by handing out dynamic assignments.

Of course I already know the answer to my question. It can be located above the '4' key on most keyboards...

Re:Business as usual

[pak9rabid]

The question I'd like to ask is why don't they just hand out a static IP by default?

There's a few good reasons off the top of my head that I can think of:

• - Dynamic IPs cut a lot of bullshit out of support calls
• - DHCP servers hand out more information that just IPs (default gateway, DNS servers, domain, etc)
• - ISPs like to have the option of re-allocating blocks of IPs without having to call hundreds/thousands of customers to have them renumber their equipment

Re:Business as usual

[XanC]

DHCP can be used to hand out a static IP.

Re:Business as usual

[Columcille]

And the good news about this is "normal people" don't have a clue what you're talking about when mentioning ipv4 or ipv6. Tell them, "We're about to run out of IP addresses!" and they'll blink at you. Tell them, "No! PANIC! IPV4 IS ALL FILLT UP!" and they will call for some friendly people to take you to a padded room. Tell them, "Oh, and we have a problem with programs that use a two digit year instead of a four digit year" and they will build disaster shelters.

Re:Business as usual

[hairyfeet]

The problem is it WILL be a disaster but not for the reasons people think of, it'll be the flyover states that cause the disaster. Frankly the pay for IT in right to work states is so damned bad nobody bothers learning more than the bare minimum and the corps sure as fuck aren't gonna pay for courses on their dime, so you have poorly trained worker with zero incentive to learn this new tech. Add to this the shitty pay and even shittier hours have caused most of the older guys like myself to run to the hills and get away from the bullshit equals a perfect storm where the old guys that could have picked it up easily are gone and the new guys hate their jobs and are doing the bare minimum until they can find another line of work.

So my prediction is while the coasts will work everything in between will rapidly fall apart. problems that would have taken an hour or two under IPV4 and which will take the same time under IPV6 on the coasts will take days or weeks simply because the skilled manpower isn't there and frankly won't be forthcoming because corps have screwed IT so badly here nobody is taking the courses anymore. My local college is thinking about dropping the IT related courses simply because everyone is in medical or law, nobody is bothering with IT as its seen as a dead end. Basically the corps made the bed but we're all gonna get to lie in it when everything that has to cross the flyover states goes to shit. Hell I called my local ISPs the other day and even their tech guys didn't know jack shit about IPV6 nor could they give me a timetable, NOT a good sign folks and talking to friends in neighboring states they are hearing the same thing.

Business opportunity

[ccguy]

Well, anyone looking to make some big bucks in the next 1-3 year should start learning IPv6. Nothing major needed, just setup a IPv6 network at home, if you can rent an external server with IPv6 in any of the many data centers that already offer it, and play with it.

It's not a lot of effort and there will be many highly paid job offers soon.

Re:Business opportunity

[SJHillman]

I think he means you can hire yourself out as an "IPv6 Changeover Consultant", spend ten minutes coming up with an IPv6 addressing scheme and then passing it on the the IT droids while taking the credit and the money.

Re:Business opportunity

[EdIII]

The problem is not learning IPv6. That's easy. At least to anyone with more than a little experience doing this. I was working before the Internet even came around and before Ethernet, so I don't see it as a big obstacle.

Where is all the fucking Enterprise hardware and firmware updates to support it?.

That's what needs to be solved. I could support IPv6 tomorrow if it was a simple firmware change. IPv6 will not be rolled out into Enterprise environments for at least 10-15 years completely. Reason why is simple. Not every network device supports it. I got clients that still have 5 years or more to go on lease contracts for huge printer and document systems. No IPv6 firmware updates in the pipeline that I know about.

Operating systems will be faster of course, but you need to cover all of the devices first.

My biggest issue is the routers themselves. If you are running a business or have branch offices, you are not, or should not, be doing that on any hardware you can pick up at BestBuy. Prosumer or higher routers that can set up multiple WAN ports don't have IPv6 yet. Perhaps the absolute newest ones might, but that could represent 20-30k in new equipment costs for a medium sized business with branch offices. For what? Just IPv6?

Unless the manufactures get off their asses, stop being greedy, and push out a firmware update for existing hardware to support IPv6 there will be a lot of people like me that have two choices:

1) Stay with IPv4
2) Spend tens of thousands of dollars on new hardware.

Tough situation.

P.S - Why do any of that until at least 1/3rd of all customers are using IPv6?

Re:Business opportunity

[klapaucjusz]

Where is all the fucking Enterprise hardware and firmware updates to support it?.

Most large companies have been requiring IPv6-capable gear for the last 4 years or so, while the DoD mandages IPv6 support since 2005.

Because of that, most recent hardware and software is IPv6-capable. Cisco IOS, for example, has been doing IPv6 since 2001. Microsoft servers have been able to work over IPv6 since Server 2003. Mac OS X since 10.4, Linux since the 2.4 series.

If you're still stuck with IPv4-only hardware or software, it's your fault.

--jch

Re:Business opportunity

[bbn]

HP did a quick one. We got a ton of IPv6 enabled HP 2910al HP layer 3 gigabit switches. Did I say layer 3? What I meant was layer 3 IPv4 and managed layer 2 IPv6 switches.

Yes, HP apparently figured the switch would qualify for all those government deals that mandates IPv6 support. I bet they were right. But the switches are just some very expensive managed switches for any IPv6 work. No support for routing IPv6 nor any other useful IPv6 support. But you _can_ telnet, ssh, SNMP etc to them using IPv6. The switch can be assigned an IPv6 address. So this qualifies for an IPv6 gold logo. They just do not advertise very loud that it is a logo for a "client"-device.

So does HP suck? Not especially so. It is the whole industry pulling this stunt.

Silly

[HBI]

Only the regional NICs have run out of blocks to distribute. No one has actually run out of IPv4 addresses. Moreover, there is a lot that still can be done to reclaim addresses. Lastly, the huge swathes of multicast and class E addresses haven't even been tapped.

This is just more attempts for the shill media to try to herd people into replacing their gear. It'll fail like the rest.

The USG was scheduled to go to IPv6 in 2006. It hasn't even begun yet.

Re:Silly

[jandrese]

In my experience the difficulty with IPv6 translation isn't at the socket layer--all of that stuff was figured out ages ago and only requires a few tweaks here and there to support both easily--the difficulty is with parsing configuration files, creating dialog boxes, etc... Lots of UI elements were spaced assuming that an IP address would only need 15 digits to be fully displayed, and IPv6 breaks that.

The upshot is that converting an application over to IPv6 is rarely as easy as it should be.

Re:Cleanup the IP Space

[SJHillman]

You assume everyone with an IP is using it to host a website. And what about people that have a redundant data link that only comes up when their main link goes down? "Well, we haven't had any downtime in the past year, guess we don't need any backups! Go ahead, take my IP!"

Re:I'll be getting ready now, thanks

[Todd Knarr]

No, I'm not a drama queen. I'm a paranoid bastard who makes Mad-Eye Moody look positively naively trusting. Which is another way of saying I've been through major infrastructure deployments before. I don't believe in Murphy, I'm on a first-name basis with the little toerag.

Making an IPv6 tunnel work, that's easy. The hard part's making it not work in the spots that need to not work without breaking what's supposed to work. If everything goes smoothly it'll be a piece of cake, and if I do it now it'll probably go smoothly. But if I wait until the last minute, 99 times out of a hundred it won't go smoothly. So I'll be paranoid and get it done now and be pleasantly surprised at the lack of problems, then kick back and relax with a bowl of popcorn while Murphy visits all the people who waited and zany hijinks ensue.

Way harder than a firmware update.

[Sycraft-fu]

For cheap consumer devices that do everything in software, sure a firmware update is all it would take, at least in theory (IPv6 can take more memory and CPU so on limited devices there might not be enough). However enterprise networking devices? They usually have to have parts replaced.

Reason is that to get the kind of speeds and latencies we want, you need ASICs, Application Specific Integrated Circuits. Those are just what they sound like: Devices designed to do a specific thing. That also means they aren't programmable. ASICs allow us to do stuff cheaper and faster than we could do in software.

A simple example is a gigabit switch. Crack one open and you see a very small little chip that handles all the switching. Now try it with a PC, stick in 8 gigabit cards and have it bridge between them. It'll overwhelm it, despite having a powerful CPU. Reason the switch can handle it is that little chip does nothing but switch packets. It is designed for only one task and does it well.

So enterprise stuff has this too, but some more complex ones. You get ASICs to speed up routing. Problem is if the ASIC was made for IPv4, it cannot be expanded to IPv6. You need a new one.

On the campus where I work they upgraded all the big routers to do IPv6 and it was pricey, seven figures even with our discounts. All the supervisor modules had to be replaced. Now yes, before that they could have technically turned it on, there was IPv6 for IOS on the older stuff. However it was all done in CPU, which is pretty limited on those routers. So if a couple people used it, it'd be fine. However if lots of people did, it'd crash the routers. The only way to give them the capacity to support it for everyone was to get new IPv6 hardware.

It isn't a matter of being greedy. As I said, Cisco would let you turn IPv6 on for many devices, like the 6500/7600s we use. It just couldn't accelerate it because it lacked the hardware. No magic fix for that.

Remember high end networking equipment isn't replaced often. You can leave it in place for over a decade. They aren't going to replace it all just for fun.

Re:Way harder than a firmware update.

[Anonymous Coward]

A PC will handle layer 2 bridging just fine /w 8 ports. There really isn't a whole lot of effort involved in shoving packets around like that. Did you base your statement on actual observation or did you pull it out of your ass?

Those desktop 5 and 8 port switches from Linksys or whatever have "switch on a chip" type IC's because they're cheap, not because they're wonderfully efficient (though obviously lower power and better latency than a PCI bus)

Obviously in a datacenter, you want a real switch.. They're more efficient, lower latency, easier to manage, and have a full array of enterprise management and configuration capabilities.

It's not YOUR addresses ...

[garry_g]

... what you don't seem to get is that the problem is not when ARIN runs out, but when your business partners get IPv6 addresses you can't reach because you didn't do your f@ckin' homework and upgrade to dual-stacked ... So go ahead, stick to IPv4, and once your boss comes in and asks why you can't exchange data with your possibly largest customer, tell him: "why would we want IPv4? Arin hasn't run out yet" ... good luck on finding a new job afterwards ...
And if you believe "Hey, no problem, it's just the Chinese and Japanese and Australians, who needs them" - think again, Europe's RIPE will run out of IPv4 addresses next ...

Re:Bologna

[evilviper]

All the articles I've read seem to come down to "it's more convenient" for applications not to have to deal with NAT... Of course it is also more convenient for people who mean to do you harm, too, since we're back to connections to outside resources coming from the machine's actual IP address, a public NATing of the private one.

NAT doesn't provide any security. Never has, never will. No, I'm not wrong. No, I'm still not wrong.

If you have a firewall between your private network and the public Internet, then you'll have all the security you want, whether using IPv4 or IPv6, with or without NAT. If you don't, then it's trivial for bad guys to reach services you don't want them to get to. If there's NAT in-between, it'll take a couple extra specially-crafted packets, but it's pretty trivial to get around.

IPv6 addresses with a firewall? Bad guys can know the IPv6 address of your valuable systems all they want, but if your firewall is blocking incoming connections by default, they can't get a single bit through to the destination.

I don't understand why people's brains turn to jello when talking about IPv6.