Comcast To Bring IPv6 To Residential US In 2010

darthcamaro writes "We all know that IPv4 address space is almost gone — but we also know that no major US carrier has yet migrated its consumer base, either. Comcast is now upping the ante a bit and has now said that they are seriously gearing up for IPv6 residential broadband deployment soon. 'Comcast plans to enter into broadband IPv6 technical trials later this year and into 2010,' Barry Tishgart, VP of Internet Services for Comcast said. 'Planning for general deployment is underway.'"

Comments

Proud to be a Comcast customer?

[csnydermvpsoft]

I never thought I'd say this, but I'm glad that I'm a Comcast customer!

(Please excuse me while I go wash out my mouth with soap)

Re:Proud to be a Comcast customer?

[tttonyyy]

Now that Comcast have bought into this whole "IPV6" thing us geeks have been falsely feeding them for years, we can all be off on the other part of the internet using our hushed-up IPV256 network (every fundamental particle in the universe needs an IP) and sniggering at their (now isolated) backwardness. ;)

Re:Proud to be a Comcast customer?

[Fallon]

I hate to say it, but I agree. As bad as all the trash talking on Comcast is, I've never had a problem. Setup was easy. The 15-20 minute call to swap out my modem for a $15 one I found at a thrift store was straight forward and easy. The only 2 real problems I had was figuring out the modem will only send out DHCP for 1 device (when you put in your firewall/router, you just need to power cycle the modem so it forgets about your PC), and the fact my dam $1,000 Cisco 1760 was the bottleneck in my network connection (replaced with a !#@$* $150 Linksys). And only one of those can remotely be called a Comcast issue.

I've never had a single connection issue in the 4 odd months I've had the service. And now I'm looking forward to messing with IPv6.

Re:Proud to be a Comcast customer?

[QuantumRiff]

Its a money saver for them. Why have a Cable TV infrastructure, and an IP Infrastructure. Think how much bandwidth they could offer if they used the entire coax connection for network. With IPv6, you make each tv channel a separate Multicast broadcast address in your network, and then the end users just subscribes to a multicast, then unsubscribes when they change channels.

Re:Proud to be a Comcast customer?

[Macrat]

I hate to say it, but I agree. As bad as all the trash talking on Comcast is, I've never had a problem. Setup was easy. The 15-20 minute call to swap out my modem for a $15 one I found at a thrift store was straight forward and easy.

Don't you consider having to make that phone call in the first place a problem?

How about their "support tools" are IE based that won't work in any browser on any platform?

And my current issue with Comcast right now is being in California and Comast routing the IP network cross country to New Jersey at 1/4 the bandwidth I had when they were routing through San Francisco.

Re:Proud to be a Comcast customer?

[David_W]

The 15-20 minute call to swap out my modem for a $15 one I found at a thrift store was straight forward and easy.

Don't you consider having to make that phone call in the first place a problem?

What's the alternative? I don't think they can just tell whose house a particular modem on their network is located in. I'd imagine they have to tie the MAC back to your account somehow, and the phone call is how they do that.

Re:Proud to be a Comcast customer?

[mmclean]

Not the original poster, but ....

I hate to say it, but I agree. As bad as all the trash talking on Comcast is, I've never had a problem. Setup was easy. The 15-20 minute call to swap out my modem for a $15 one I found at a thrift store was straight forward and easy.

Don't you consider having to make that phone call in the first place a problem?

Nope, not in the slightest. I would expect to have to call to initialize service - as I do for cell phones, pizza delivery, etc.

How about their "support tools" are IE based that won't work in any browser on any platform?

Never need to use them, never bothered installing them. I've seen from someone else's connection what they are and it's nothing I can't do on my own without them (ping, traceroute, search comcast.com help files, etc.). Those Comcast tools exist for the non-Slashdot crowd.

And my current issue with Comcast right now is being in California and Comast routing the IP network cross country to New Jersey at 1/4 the bandwidth I had when they were routing through San Francisco.

My Jacksonville connection routes through Atlanta, my NJ connection yrs ago routed through somewhere in NY/NJ IIRC. My bandwidth has been fine in both places.

Looking forward to IP6 also (though I'll have to get rid of my $100 cheap router for a "real" one)

Re:Proud to be a Comcast customer?

[jmilne]

Because there's no such thing as IPv4 multicast... Oh, wait. That's exactly what cable companies have already been doing with switched digital. Multicast isn't the main reason a cable company would go with IPv6. The biggest problem Comcast (and other cable companies) has is that your cable modem gets two, and sometimes three IP addresses, let alone all those set-top boxes doing that switched digital. One to manage it, one to give you your "public" IP, and perhaps a third for your phone. 24 bits (10.0.0.0/8) only gives you 16 million addresses, and that's assuming you're utilizing them rather effectively. They're probably using the 172.16.0.0/12 for their internal network, but even so, that only gets you an extra million addresses. Look at the number of customers Comcast has, and you begin to see the problem they have just with addressing all those cable modems and set-top boxes.

Don't expect to be getting your own IPv6 address any time soon. Most likely, they're going to roll it out for managing all those devices first, and you'll still be assigned an IPv4 address for your Internet connectivity.

Re:Proud to be a Comcast customer?

[Andy Dodd]

I believe this already exists for cable systems w/o IP Multicast - It's called switched digital video.

That said, if they implemented multicast out to and through the backbone, it could save a LOT of upstream bandwidth from user P2P apps.

Imagine if all subscribers to a torrent could receive multicast from the seeder, as opposed to now where the seeder gives peers content and they forward it on. Most P2P is effectively "ghetto multicast", with lack of backbone participation severely reducing efficiency.

Re:Proud to be a Comcast customer?

[Danathar]

You've got that right!

May NAT die a horrible and torturous death.

Re:Proud to be a Comcast customer?

[RoFLKOPTr]

Don't you consider having to make that phone call in the first place a problem?

How does that even make sense? Of course you're gonna have to call them... they don't give away internet for free, so yes you have to talk to them to have your modem registered to the account so that the DHCP server will talk to you. I think it's awesome that they let him use a third-party modem at all.

Frankly, I love Comcast. Yeah, they have a tendency to be jewish, but they're a huge corporation. Yeah, they have outages, but who the hell doesn't? Yeah, they were throttling peoples' torrents, but I never experienced them throttling mine so I don't really give a damn. I've never had any real problems with Comcast. Their outsourced phone support sucks, but you only get that about half the time, so you just have to ask the person who gives him a paycheck and tell him to transfer you to a Comcast call center.

I really don't see why so many people have a problem with Comcast. And I mean, problems with Comcast as a company, not with regards to your stupid commie philosophy big corporations are bad bullshit, because we are a nation built upon Capitalism, and we always will be, and even with all the problems of Capitalism it's way better than Communism (see also Russia).

Re:Proud to be a Comcast customer?

[langelgjm]

FWIW, I got their business-class internet and have been pretty happy with it. You pay a small premium over the consumer-oriented service (no 6 month introductory rate, and $17 / mo higher than the standard consumer rate), but they specifically told me there's no cap (and I haven't had any issues with that). Customer service is also separate from home users, which is great - short hold times, when I once had a problem, they sent someone out the next morning to fix it.

Re:Proud to be a Comcast customer?

[Anonymous Coward]

Not me. I have a 5-minute to 1-hour outage about once a month. I also have funny DNS issues, and sometimes routing issues, that usually last from 5 mintues to an hour. One other thing: I have UPSs for my home computers - but Comcast doesn't have UPSs or generators for any of their network equipment (at least in my area), so power outages kill my connection even though my equipment stays up.

I'd switch to DSL in a heartbeat.

Re:Proud to be a Comcast customer?

[Lumpy]

?Which internal network.

They have tons of them.
1 for ad-sales
1 for cablemodems
1 for cable boxes
1 for corperate office, well actually 390 of those.
  Ad sales offices have their own
Cable operations has their own
Corperate has their own
On demand has their own
It has their own.

Jeebus, if I dug out my old documentation I'd bet I have nearly 20 pages of networks used at comcast as of 4 years ago.

What they want IPv6 for is for when they force cable boxes on all of you. The new digital boxes collect granular data on your viewing habits so they can collect that and sell it. Realtime demographics makes Nielsen data a joke.

Plus they can fire 90% of their installers, if cable is live to every home and they simply turn off the box, they can increase profits and reduce jobs. They are fighting HARD to get rid of any cable signals that are analog or unencrypted.

Re:Proud to be a Comcast customer?

[Macrat]

How does that even make sense?

Oh, like maybe a web form where you enter the MAC address of the new router before you upgrade?

Or in the case of new service you can give that info when you order instead of having to deal with a clueless service tech that only knows how to run an app on a Windows CD?

There are lots of things that Comcast makes more difficult than it needs to be.

Re:Proud to be a Comcast customer?

[DragonWriter]

What's the alternative? I don't think they can just tell whose house a particular modem on their network is located in.

Why not? Maybe there is some fundamental difference between DSL and Cable infrastructure that prevents this, but when I've switch modems on my DSL line I've just unplugged the old one and plugged in the new one.

Re:Proud to be a Comcast customer?

[RoFLKOPTr]

There are lots of things that Comcast makes more difficult than it needs to be.

True... but is it really THAT big of a deal? No. Besides, there's the possibility that there's something a little more technical going on than just entering a MAC Address. Oh and there's also the enormous possibility that some fuck will buy his own modem for whatever reason and then call Comcast when it doesn't work right when he programs it and bitch about "your damn little web system isn't working right."

Re:Proud to be a Comcast customer?

[Just Some Guy]

Nope, not in the slightest. I would expect to have to call to initialize service - as I do for cell phones, pizza delivery, etc.

Really? That's kind of strange. When my DSL modem went south, I popped in a spare and configured it like the original and was back online without a single call to my ISP.

Re:Proud to be a Comcast customer?

[Dog-Cow]

That's because the line going to your house from the Telco is unique; it only goes to one place. It's the same reason you can add a jack for POTS and plug in any old phone and expect it to work.

Re:Proud to be a Comcast customer?

[amorsen]

Why couldn't you make an IPv4 multicast address for each TV channel?

Re:Proud to be a Comcast customer?

[Tacvek]

Yes there is a fundamental difference. In DSL you have an individual line to the the phone company owned equipment (the DSLAM). Thus any data on that line is either data from you or data intended for you. On cable, your neighborhood shares a line. That is to say, that on the cable line that comes into your house is not only your data, but the data of the people next door (if they use the same cable internet service.) To prevent you from seeing the neigbors data, and to determine who sent anything in the other direction, the data is encoded (I would not dare call it encrpyted) with a modem specific identifier.

Re:Proud to be a Comcast customer?

[Tacvek]

Indeed. I am always shocked that people install any software provided by the ISP. They don't need some broken net-nanny software, a half-assed firewall that does not work as well as the Windows firewall it disables, having branding adding to internet explorer and outlook express, yet another worthless IE toolbar, or even worse, some form of stand-alone (screen edge docking) toolbar. The only feature that might be reasonable, is changing the IE homepage, and they can install software to do that.

Then again, I also find it incredibly annoying that home routers come with big warning stickers that you should install the software on the cd, since that software is in no way nessisary thanks to the web interface. Indeed, I honestly have no idea what is even on said CDs, but whatever it is, I don't miss it.

Re:Proud to be a Comcast customer?

[Ucklak]

Maybe I'm not reading into this deep enough but if they give you an IPv6 address which doesn't do NAT, can't they also say 5 devices max?
The 5 devices limit is in the fine print for current connectivity even though we all use routers to give us more.

I can see how this can be used to nickel and dime the customer as they do now with 'tuners' and 'cablecards'.

I guess we could still use IPV4 routers to talk to the IPV6 connection.

Re:Proud to be a Comcast customer?

[vtcodger]

***Don't expect to be getting your own IPv6 address any time soon. Most likely, they're going to roll it out for managing all those devices first, and you'll still be assigned an IPv4 address for your Internet connectivity.***

You're probably right. Too bad in a way. I expect that there will be entertainment galore starting shortly after early adopters turn off their NAT routing. Maybe it's just me, but I sort of think that a large part of Internet security such as it is probably depends on the limitations of IPv4 and NAT. With IPv6, my computer, phone, TV setup box, refrigerator, printer will all be fully visible and accessible to any corporation, government, nutcase, religious fanatic, scam artist, and sociopathic teenager anywhere in the world? What could possibly go wrong?

Re:Proud to be a Comcast customer?

[Randle_Revar]

Don't expect to be getting your own IPv6 address any time soon. Most likely, they're going to roll it out for managing all those devices first, and you'll still be assigned an IPv4 address for your Internet connectivity.

They have been talking about IPv6 for management for years, they just needed to wait for DOCSIS 3.0 to support it. But this article specifically talks about residential, consumer IPv6.

From the article:

With wholesale availability of IPv6 on the Comcast network, the next steps include making IPv6 available to Comcast's business and residential broadband customers.

"Comcast plans to enter into broadband IPv6 technical trials later this year and into 2010," Tishgart said. "Planning for general deployment is underway."

Re:Proud to be a Comcast customer?

[Anonymous Coward]

The 15-20 minute call to swap out my modem for a $15 one I found at a thrift store was straight forward and easy.

Don't you consider having to make that phone call in the first place a problem?

What's the alternative? I don't think they can just tell whose house a particular modem on their network is located in. I'd imagine they have to tie the MAC back to your account somehow, and the phone call is how they do that.

Yup, exactly correct. (I work at a cable-modem based ISP) The other replies below are also correct in the difference between the dedicated DSL line and the "shared" coax lines for cable modem plant.

I will throw one other thing in the mix- if a cable modem is provisioned, it can actually be used at any hot outlet which runs off the same headend or CMTS equipment. So as long as the cable modem's MAC is active in their system, you can actually unhook it, and take it to a buddy's house, for example.

There are a lot more details that I won't go into, but the answer is that when using a cable modem technology, the ISP has to have the MAC address of each device on the system registered. When the modem connects, it downloads a config file that sets up your speeds and a few other things, so your MAC is tied in with what speed package you pay for, and that's how the system knows which one to send to your modem. In addition, the specific config file often has to have make & model specific information, so for example the binfile for a motorola isn't necessarily the same as one for a linksys, etc.

And no, this is not unique to Comcast in case you hadn't figured that out yet, this is true for any ISP that uses cable modems.

Re:Proud to be a Comcast customer?

[Anonymous Coward]

They already have rolled out IPv6 to manage customer devices because of precisely the problem you describe.

Re:Proud to be a Comcast customer?

[Anonymous Coward]

Most ISP's and other large networks will run the cable modem's internal IP on the 10. scope. Big companies like Comcast will actually have their network segmented into regional groups, instead of trying to run everyone on one massive network. Any time they might start running short on 10.'s in an area they can just sub-segment it as needed. Of course that is a very simplistic explanation, it get's a lot more in-depth in practice. Generally each customer will have a 10. for the cable modem, and if they have phone they will have a second IP for the telephony portion.

The issue is really a matter of the public customer equipment IP's. In order for their routing to work properly, they have to assign sub-blocks of their IP pool into their regional markets, so they can end up in situations where one market has a lot of unused IP's and another is running short. See my first point, it ties in directly with this as well.

Don't expect to be getting your own IPv6 address any time soon. Most likely, they're going to roll it out for managing all those devices first, and you'll still be assigned an IPv4 address for your Internet connectivity.

I agree. They will most likely not do a system-wide launch either, they will probably add a few test markets first. There is a lot of customer equipment, especially routers, which can't yet handle ipv6 properly. And don't say use a custom firmware, we're talking about large numbers of technology illiterate people here.
My guess is that they'll offer the ipv6 option for those who wish to use it, and will hazard a guess that they won't force anyone to migrate off ipv4 until at least 2015, and would not be surprised to see them still running ipv4 in 2020.

Re:Proud to be a Comcast customer?

[Anonymous Coward]

And if you do get them, you'll wish you had stayed with NAT, because they'll probably charge if you have more than (insert low number here) computers/devices that need IP adresses. Finally a way for them to get more money out of the geeks!

Re:Proud to be a Comcast customer?

[Drinking Bleach]

Looking forward to IP6 also (though I'll have to get rid of my $100 cheap router for a "real" one)

Or upgrade to OpenWrt (if it supports your router), and you might not need to get rid of anything. I have a $50 cheap router with full IPv6 capability.

Re:Proud to be a Comcast customer?

[Imagix]

It could be encrypted as well. See Baseline Privacy Interface (BPI, or BPI+).

Re:Proud to be a Comcast customer?

[sjames]

The one issue I have with Comcast is that their call center thinks every problem means you need to replace your modem.

The uptime is fairly decent. I haven't seen any need for "support tools". Everything I need to know I can get from my cable modem's web interface (mostly the system log) or from the usual ping, mtr, etc on my (Linux) PC.

I can't comment on your routing, I'm on the other side of the country. The routing here is fairly sane.

Re:Proud to be a Comcast customer?

[zonky]

There are private addresses in IPv6. That said, NAT is _not_ a form of security.

Re:Proud to be a Comcast customer?

[Chabo]

And my current issue with Comcast right now is being in California and Comast routing the IP network cross country to New Jersey at 1/4 the bandwidth I had when they were routing through San Francisco.

They were re-routing your traffic so the NSA wouldn't be able to read your e-mail. :)

Re:Proud to be a Comcast customer?

[tagno25]

Looking forward to IP6 also (though I'll have to get rid of my $100 cheap router for a "real" one)

My router was only $100 plus $50 shipping(4U computer) plus $40 for the OS, but could have been only $100 and done the same in a smaller space(routerboard.com). It has BGP, OSPF, RIP, IPv6, Bandwidth Queues, NAT, and more.

Re:Proud to be a Comcast customer?

[ls671]

Well could this indicate that if they do this, someone (like linux kernel developers?) will come up with a way to masquerade several internal LAN ipv6 addresses as one ipv6 address ?

Re:Proud to be a Comcast customer?

[Anonymous Coward]

How about their "support tools" are IE based that won't work in any browser on any platform?

What are you talking about? I had Comcast for 4 years and I don't even own any Windows machines. The biggest complaint I had, the one time I had a service problem, was that I had to explain to the tech support guy what OpenBSD is. Even that wasn't too big a deal, as "It's kinda like Linux" was all he needed to know.

Re:Proud to be a Comcast customer?

[Anonymous Coward]

hm, nice idea, but how would they call this... maybe NAT?

Re:Proud to be a Comcast customer?

[csnydermvpsoft]

The funny thing is, after posting this, our connection (a $100/month Comcast Business line) was down for the majority of the afternoon. Their support was very responsive, though.

Re:Proud to be a Comcast customer?

[Aqualung812]

they have a tendency to be jewish

About the same tendency that you have to be a racist. That crap needs to stop.

Re:Proud to be a Comcast customer?

[Lumpy]

They cant stop NAT. I can always NAT no matter what they try.

That's the cool part, they really cant set a "you cant nat" bit, because I can always ignore it (OpenWRT will never bow to them)

Plus if you NAT right, they cant even detect it. Prove that I'm not running 40 different programs on my single PC.

They tried that crap at MSU when I went there. I fixed their little red wagon and had 4 machines connected to my 1 pc connection. And yes you can ipv6 NAT to an ipv6 network I've done it in testing, and it's not really that hard.

Re:Proud to be a Comcast customer?

[ls671]

Masquerading is a specific subset of NATing.

Masquerading is included in NAT.

Masquerading != NAT, stricly speaking.

Masquerading is a more specific term, thus more precise than NAT ;-)

There is much more stuff that you can do with NAT that is unrelated to masquerading several addresses as one.

Asprin

[Kid Zero]

Do they make enough painkillers to deal with the headaches this'll cause?

Otherwise: Good Luck, guys! You'll need it.

Re:Asprin

[Techman83]

Meh, good on 'em. Gotta start some time! The longer we leave it, the worse it will get. IPv6 isn't really a big deal at a protocol level, it's just all the stuff that isn't IPv4 ready and IPv6 -> IPv4 tunnel or Dual Stack will sort that out...

Re:Asprin

[Shakrai]

Do they make enough painkillers to deal with the headaches this'll cause?

Maybe somebody told them that IPV6 makes it easier to inject fake RST packets into TCP connections ;)

Re:Asprin

[swillden]

Do they make enough painkillers to deal with the headaches this'll cause?

What headaches are those? Have you dealt with IPv6 at all? It's very easy to work with, and co-exists perfectly well with IPv4. I set up IPv6 in my house with a tunnel and it was amazing how smooth it was. I set up the IPv6 tunnel and addresses on my router (that was a little tricky -- but no more than any other router configuration), started up radvd, which periodically broadcasts an announcement about what the local IPv6 router is, and instantly every machine on the network -- Linux, Mac and Windows -- had an IPv6 address in addition to their private IPv4 address (10.x.x.x). Of course, the typical home user couldn't do any of that stuff, but they don't have to if the v6 service comes directly from their ISP.

What's more, I was surprised to note that as soon as all my computers had v6 adresses, they started using them! IPv6 DNS is in place, and all decent applications do an IPv6 name lookup in parallel with the IPv4, and if they get an IPv6 answer, they connect via v6. I know Firefox does because I have a Firefox add-on that shows the IP of the web server in the status bar, and sometimes I come across sites for which it shows a v6 address.

About the only part of the infrastructure that really isn't ready, as far as I can tell, is everyone's home routers. Those ubiquitous Linksys boxes mostly don't support v6 unless you put third-party firmware on them (which I did, but most people obviously wouldn't do). But I'm sure the next generation or two of home routers will come with IPv6 support enabled and it will Just Work. Oh, and they'll also be configured by default to reject externally-originated connections, so that Joe Sixpack will still have the same level of firewalling he has with NAT -- but with lower overhead and fewer limitations. Until those routers are widely available, v6 and v4 can coexist quite nicely.

I predict that this will be relatively painless for Comcast's techs, and completely transparent to their customers.

Re:Asprin

[Randle_Revar]

I have doubts about it causing many problems for users. Vista and Win7 have IPv6 turned on out of the box. If the user one of those OSs, and has a DOCSIS 3 modem attached directly to the computer, or they have a IPv6 router, they should be assigned an IPv4 and an IPv6 address. If the OS or the router doesn't support IPv6, they would just get an IPv4 address.

Mac and Linux of course, both support IPv6. Even setting up IPv6 manually with a tunnel isn't that hard, if you know just a bit about networking.

Of course nothing ever goes as smoothly as it "should", but I bet more of the issues will be on the Comcast side.

Re:Asprin

[mdmkolbe]

Many users may already be using IPv6 without knowing it. When my ISP added IPv6 support, it was so smooth that I didn't even know they had added it until one day I noticed the extra entry in my ipconfig. Check http://www.whatismyipv6.net/ to see if you have an IPv6 address.

Re:Asprin

[Anonymous Coward]

How did all of your IPv6 nodes get the address of your DNS server to lookup IPv6 addresses? DHCP over IPv4? Thought so.

So, exactly what benefit have you achieved by configuring IPv6 on your network? It is still useless without IPv4. And if you need IPv4, you are still bound by IPv4's limitations.

Re:Asprin

[swillden]

How did all of your IPv6 nodes get the address of your DNS server to lookup IPv6 addresses? DHCP over IPv4?

I've done it both ways. It's not hard to advertise an IPv6 DNS server with radvd and run rdnssd on the client machines to see the advertisements and configure resolv.conf. No DHCPv4 required.

However, at present I'm using IPv4 for DNS. DNS works fine via either v4 or v6, and whichever protocol you're using to query the DNS server, you can ask for A or AAAA records, or both. They coexist quite nicely.

Thought so

Thought so what? I'm not sure why you think using a mixture of v4 and v6 eliminates benefits of v6.

So, exactly what benefit have you achieved by configuring IPv6 on your network? It is still useless without IPv4. And if you need IPv4, you are still bound by IPv4's limitations.

What benefit have *I* gained? It was an interesting experiment, and I do now have fixed, publicly-routable IP addresses for all of the machines in my house, even though my ISP gives me a single, dynamic (though rarely-changing) IP. I've made use of that a few times while traveling, to SSH directly into machines at home. It's a minor convenience, since I can always SSH to my public IPv4 address, and then SSH from there into specific machines. It's nice to avoid the extra step, though.

And you're wrong about being bound by IPv4's limitations. I'm not. IPv6 works just fine through the tunnel. What I am bound by, though, is the limited adoption of IPv6. For example, there are bittorrent trackers that operate on IPv6, so I can run a BT client on a machine behind my firewall without having to set up port forwarding -- but only other peers on IPv6 can connect to me. If and when the set of available peers goes from "geeks who mess with IPv6 tunneling" to "people with Comcast", then that will instantly become very useful.

Although few web sites, etc., are available via IPv6 as of yet, I do notice another interesting advantage -- speed. There is a small but noticeable improvement in latency going through the tunnel. I don't know if it has to do with avoiding NAT overhead, or something to do with the routes or what, but it's noticeably faster. Or maybe TCP over IPv6 is different and has faster handshake/startup times? Dunno. I'm puzzled by that difference, and don't know if it will still hold when I get IPv6 from my ISP (which is Comcast), but I'm hopeful.

It's Comcastic

[slashtivus]

I have Comcast. Typing ipconfig into my command prompt returns IPV6 addresses.

I did not RTFA but it seems to me that they have already started with this in 2009.

Re:It's Comcastic

[Anonymous Coward]

For those in California, Surewest has been doing this, on their fiber at least for the last year or two, maybe longer. I haven't gotten much use out of it since I'm still running an older Linksys router which can only pull IPv4 addresses, but when I hooked my computer up directly a few months ago it was giving out IPv6 addresses via DHCPv6.

Re:It's Comcastic

[quazee]

Are you sure these are not 6to4 addresses (2002:::xxx)?
By default, Vista and Win7 will automatically allocate a 6to4 address for each non-private IPv4 address configured on the computer.
(since you mentioned ipconfig and not ifconfig, I assume you are using Windows)

Re:It's Comcastic

[bjackson1]

Are you directly on Comcast or are you behind a router?

I have a WRT54G running Tomato and Comcast gives it a IPv4, and Tomato assigns IPv6 to my internal network.

Re:It's Comcastic

[dascritch]

Saying hi from 2a01:e35:2f1e:a290:21a:92ff:feb8:bfa8/64
thank you free.Fr

Re:It's Comcastic

[jonfr]

Try going to Sixxs.net (IPv6 ready) and see if you connect with a Global IPv6 address or not. Local-Link IPv6 is just your standard IPv6 that comes with Windows XP/Vista, it is not good for anything that I know of.

Re:It's Comcastic

[Anonymous Coward]

Vista has Teredo adapters that allow IPV6 over IPV4

Re:It's Comcastic

[Anonymous Coward]

Much to my surprise when I visited that site I was greeted with the message "You've got IPv6!"

For the record, I'm a Comcast user in Jacksonville, Florida. And now that I think of it, not so long ago my connection when out and when I called they said something about an upgrade.

Re:It's Comcastic

[Anonymous Coward]

If the machine is internet-facing and the IPv6 address starts with 2002, that IP was automatically configured. If your IPv6 starts with 2001:, then it's most likely Teredo, (assumed-- since you said you didn't configure anything..either way it has nothing to do with Comcast and everything to do with IPv6 technologies in the Windows operating system. Thanks

Re:It's Comcastic

[Anonymous Coward]

I just pulled this from my Linux server, which sits directly behind the Comcast Cable modem. (redacted where appropriate)

$ ifconfig
Link encap:Ethernet HWaddr 00:**:**:**:**:**
inet addr:67.***.***.** Bcast:255.255.255.255 Mask:255.255.240.0
inet6 addr: fe80::***:****:****:****/64 Scope:Link

Re:It's Comcastic

[XanC]

That's a link-local address. It doesn't do anything for you in the wider world.

Re:It's Comcastic

[swillden]

inet6 addr: fe80::***:****:****:****/64 Scope:Link

No need to redact that. It's a link-local, non-routable address, not usable by any machine not directly connected to your LAN. You don't have IPv6 service.

Re:It's Comcastic

[caluml]

Lots of people forget to firewall IPv6 too. It makes for interesting snooping.

$ telnet 2a01:e35:2f1e:a290:21a:92ff:feb8:bfa8 22
Trying 2a01:e35:2f1e:a290:21a:92ff:feb8:bfa8...
Connected to 2a01:e35:2f1e:a290:21a:92ff:feb8:bfa8.
Escape character is '^]'.
SSH-2.0-OpenSSH_5.1

Protocol mismatch.
Connection closed by foreign host.
$

I bet your v4 tcp/22 is firewalled though.

Re:It's Comcastic

[XanC]

Um... That address does start with fe80.

Re:It's Comcastic

[Anonymous Coward]

Tomato doesn't support IPv6 yet, genius.

Re:It's Comcastic

[spydabyte]

Are you connected directly to the ISP pipe or do you have a home router? Oh and the title of the post is misleading. They're going to have trials in 2010.

Re:It's Comcastic

[berashith]

nope .. that starts with "inet6 addr:"

sheesh ... learn to read

 

Re:It's Comcastic

[uid8472]

inet6 addr: fe80::***:****:****:****/64 Scope:Link

No need to redact that. It's a link-local, non-routable address []

The lower 64 bits almost certainly contain the interface's MAC address; while it's not as bad as a globally reachable network address, some people still might not want to post it openly on /.

Re:It's Comcastic

[slashtivus]

I appear to have both.
My "Local Link IPv6" starts with "fe80::34cb:18f0" etc , that does not appear to be a 2002 address.
However,
I also appear to have a tunnel adapter that starts with 2002 as you mention.

I'm using Vista, I am connected directly to my cable modem PDX area.

Thanks for the reply I might have learned something new today. I'm not a networking person. Does that mean I'm still in translation mode?

I also have an IPv4 address listed below the IPv6.

Re:It's Comcastic

[ion.simon.c]

Explain to me again the risk inherent in exposing a strictly key-based-auth SSH daemon to the Greater Internet?

Re:It's Comcastic

[Danathar]

If you don't have Teredo tunneling turned on you will get what is called a Link Local scoped IPv6 address. If Teredo is on then you are on a 6to4 tunnel.

Good news..

[Manip]

That's great news for the people within the trial area. They will have much more free time to, you know, go out and meet women. Since now a ton of web-sites break when they attempt to visit them.

If it was just a matter of software updates, but alas there are mountains of sites that are literally hard-coded to store IPv4 addresses and you get a nice PHP error when you attempt to visit them.

IPv6 is the new Y2K.

Re:Good news..

[vivimage]

Ipv6 would run as a dual stack so unless your retarded and set AAA records or ipv6 access there will be ZERO PROBLEMS if you do you suffer from a layer 8 problem

Re:Good news..

[Just Some Guy]

If it was just a matter of software updates, but alas there are mountains of sites that are literally hard-coded to store IPv4 addresses and you get a nice PHP error when you attempt to visit them.

I guess I live a sheltered life, because I've been using IPv4 and IPv6 in parallel for about 7 years and I've never had a site break like that.

Re:Good news..

[Anonymous Coward]

Could be layer 9, too.

Re:Good news..

[Randle_Revar]

That doesn't make sense. If you have IPv6, DNS will first look for an IPv6 address for a given name. If it is found, it is used. If the name cannot be resolved to an IPv6 address, IPv4 will be checked. If a v4 address is found, it is used. If it does not resolve, you get the usual error.

The only way for something like what you describe to happen is if v6 DNS was setup, but the http server was not setup for IPv6.

Re:Good news..

[Anonymous Coward]

Maybe we're both lucky. I'm behind a Comcast router using IPv4 and IPv6 in parallel for almost five years without a hitch.

Re:Good news..

[sjames]

Only if the service is v6 only (doubtful for a trial) or the site is stupid enough to publish a bogus AAAA record.

As a user, what do I care?

[BadAnalogyGuy]

As long as DNS works fine, and I can access all my favorite porn sites, I don't care what is going on under the covers.

For all I know, it could be hamsters squeaking in HyperCard. As a user, it really doesn't matter.

Re:As a user, what do I care?

[berashith]

You should care what is going on under the covers at porn sites. The point is really to not have covers in the way in the first place, unless that is what you are really into.

You've got the protcol

[sakdoctor]

Now buy the T-shirt.
There's no place like ::1 (0:0:0:0:0:0:0:1)

Re:You've got the protcol

[Anonymous Coward]

There's no place like localhost?

Re:You've got the protcol

[Palshife]

You. Turn in your geek card. Now.

Re:You've got the protcol

[Chris Mattern]

Oh, for heaven's sake.

Just say, "There's no place like localhost."

You have an /etc/hosts file for a *reason*, you know.

Re:You've got the protcol

[Anonymous Coward]

Surely unless it's "There's no place like ~" then you and the OP should turn in their cards.

Re:You've got the protcol

[Xtifr]

Just say, "There's no place like localhost."

Especially since that's what the popular t-shirt really translates to in the first place, which is why I've never bought one; it makes no sense. On the other hand, if they offered "there's no place like ~", I'd snap that up in an instant! :)

When I ssh into home from work, the IP addy I use bears no resemblance to 127.0.0.1, nor have I ever heard anyone, admin or developer, refer to localhost as "home".

Re:You've got the protcol

[Frosty Piss]

"~ Sweet ~"?

Re:You've got the protcol

[Palshife]

http://www.thinkgeek.com/tshirts-apparel/unisex/generic/5d6a/

Worn by millions. Get one before they're obsolete :)

Re:You've got the protcol

[Anonymous Coward]

That might have other implications ~~~x_o

Re:You've got the protcol

[swillden]

::1 is shorter

And cooler.

what about caps?

[gandhi_2]

Will comcast unveil a "tiered plan" whereby you only get the first 5 groups of four hexadecimal digits at the base price, with prices increasing up to 8?

Re:what about caps?

[spydabyte]

That'd technically be really nice, but expensive. If they could give me a unique net in the top 8? I'd be golden.

I still don't like IPv6

[tjstork]

IPv6 is like the phone company saying, hey, we have a (aaa) eee-nnnn system doesn't have enough room, so let's replace it with a system that has 20 digits.

It just sucks to use for consumers, making everyone else's life more complicated just to simplify it for the service providers.

I would prefer an addressing system that simplifies life for me.

Re:I still don't like IPv6

[Timothy Brownawell]

IPv6 is like the phone company saying, hey, we have a (aaa) eee-nnnn system doesn't have enough room, so let's replace it with a system that has 20 digits.

It just sucks to use for consumers, making everyone else's life more complicated just to simplify it for the service providers.

I would prefer an addressing system that simplifies life for me.

What it's supposed to mean is that every computer can have a public address. So if you sign up with one of the dynamic DNS providers (which will probably be integrated with your OS fairly soon) you should be able to share pictures and things from your own computer without having to upload them to somewhere, or be able to log in remotely to look at some file (private) you forgot to bring with you, or any number of other things (fewer firewall errors on p2p networks? true p2p voip, without needing to sign up with a service that lets you punch holes in NAT?). This would also work without the dynamic DNS provider, but the URL would look uglier.

Most likely, this would also lead to relaxing the typical rule ISPs tend to have against running servers on home connections. They can't really forbid something that gets built into the OS like these sorts of features probably will.

Re:I still don't like IPv6

[MaerD]

It's slightly worse. It's more like the phone company going "we can only handle phone numbers from (000)000-0000 to (255)255-2555" and instead of going "Hey.. let's try making go up to (999) 999-9999 and maintain the pattern everyone knows, or even say adding another set of numbers to make 255(255)255-2555 available, let's change it all up into some long string people can only half pronounce and you have to be a telephone repairman to understand... your new phone number is now ab823:fff::324223 and your neighbor is ab823:fff:731:823:324223". Can you imagine the confusion?

I never liked ipv6 is you end up with addresses like 2001:0db8:85a3:0000:0000:8a2e:0370:7334 that can also be written as 2001:db8:85a3::8a2e:370:7334. Trying to get "what's your ip address?" when doing telephone technical support is going to be nightmarish. Not just from the fact it's now a long hex string, but also from a complete lack of understanding by users, much less some level 1's I've dealt with.

Heck, just try diagnosing a user who "can't get to the internet" and it turns out to be a wrong dns server entry. It's hard enough to get them to go to google's ip now.

Re:I still don't like IPv6

[Leebert]

It's more than that. For example, a big part of ipv6 is trading off some degree of address portability for routing efficiency. And stateless autoconfiguration. And ipsec. Address deprecation. Mobile ipv6.

There's lots of advantages. (Granted a few of the advantages end up being disadvantages...)

Re:I still don't like IPv6

[oldspewey]

I would prefer an addressing system that simplifies life for me

Agreed. What I'd really like to see is some kind of naming protocol so I don't have to remember all these long strings of numbers separated by dots. It would be awesome if internet addresses were identified by an alphanumeric name, then when I use that name there is a server somewhere that figures out what IP address that name is really pointing to.

I bet if everyone here at 216.34.181.45 put their minds to it we could even come up with something here.

Re:I still don't like IPv6

[characterZer0]

They can't really forbid something that gets built into the OS like these sorts of features probably will.

Of course they can, and they will.

Re:I still don't like IPv6

[TuaAmin13]

Crap. That was one of those slashdot comments that don't really require a response.

Feel free to whoosh! me.

Re:I still don't like IPv6

[edmicman]

So instead of going to flickr I have to know and maintain all of my friends' computer addresses? In what, an address book that I store on my computer? What if I'm at a friend's house and want to show them another friend's picture, but they don't know the address?

I agree there are good reasons to go to v6, but directly accessing every device via a public address is not the answer, unless it's made really REALLY transparent and easy to use. Who's going to manage that? The OS?

Re:I still don't like IPv6

[Anonymous Coward]

Windows will do it through the Windows Live ID. They're already geared towards it. They want to manage your everything.

Re:I still don't like IPv6

[Timothy Brownawell]

They can't really forbid something that gets built into the OS like these sorts of features probably will.

Of course they can, and they will.

Sure, they can... just like they could and did RST your bittorrent connections, or throttle/cap traffic to services that compete with their services.

Until people get pissed because they now know what's being taken away, and maybe get congress or the FCC or FTC involved.

Re:I still don't like IPv6

[Just Some Guy]

IPv6 is like the phone company saying, hey, we have a (aaa) eee-nnnn system doesn't have enough room, so let's replace it with a system that has 20 digits.

How often do you enter IP addresses directly?

It just sucks to use for consumers, making everyone else's life more complicated just to simplify it for the service providers.

How so? I'd be surprised if most consumers ever noticed.

Re:I still don't like IPv6

[silent_artichoke]

That was one of those slashdot comments that don't really require a whoosh.

Re:I still don't like IPv6

[slimjim8094]

No, it's like saying there's 6 billion people in the world and 000-000-0000 through 999-999-9999 isn't enough to fit every telephone, fax, and modem in the world, so let's make it longer. At the same time, instead of dialing this crazy long phone number we'll never run out of digits for, you just pick up the dial and say "Joe Moriarity down the street, second floor bedroom" or "Comcast billing" - both of which have no lasting relation to the actual digits.

Not nearly so bad as you make out. A string of digits doesn't mean anything to humans, who you're calling does.

Re:I still don't like IPv6

[SolarCanine]

Who's going to manage that? The OS?

Meet my good friend DNS:

Hey, Joe - check out the pics that Tonya just put online: tonyas-pc.smithhousehold.comcast.net/pics/

Just because IPV6 is coming into play doesn't mean we suddenly jettison DNS - let's not go creating problems where they don't exist...devices have been self-registering in DNS via DHCP for a looooooong time - hell, even Microsoft OSes do it ;)

Re:I still don't like IPv6

[tlhIngan]

What it's supposed to mean is that every computer can have a public address. So if you sign up with one of the dynamic DNS providers (which will probably be integrated with your OS fairly soon) you should be able to share pictures and things from your own computer without having to upload them to somewhere, or be able to log in remotely to look at some file (private) you forgot to bring with you, or any number of other things (fewer firewall errors on p2p networks? true p2p voip, without needing to sign up with a service that lets you punch holes in NAT?). This would also work without the dynamic DNS provider, but the URL would look uglier.

Most likely, this would also lead to relaxing the typical rule ISPs tend to have against running servers on home connections. They can't really forbid something that gets built into the OS like these sorts of features probably will.

No, it'll be an excuse for an ISP to give you a /64, but firewall out all but the number of addresses you get unless you pay for more.

And servers will still be banned - there's not enough bandwidth upstream from most connections to handle everyone serving something (last mile problem).

Everyone thinks IPv6 is the magic savior - it'll enforce net neutrality, it'll prevent your PC from getting infected, it'll solve the public IP issue, it'll solve NAT issues, it'll have QoS for real, blah blah blah.

Sure IPv6 has it all. But I doubt any ISP will do business any differently with IPv6 than otherwise. In fact, they'll just salivate that any caps will be reached a bit quicker because of the increased IPv6 header size. Mobile operators are probably salivating as well - 5 cents per kilobyte (not kiB), which includes the OTA headers, plus increased IPv6 header size, means the real payload per packet goes down, and more data usage results (== more $$$ - the incremental network cost for IPv6 is low to the network to support IPv6, but not you the user have to pay more for the same traffic since the amount of data you need to transfer increased).

I see IPv6 as allowing an ISP to ding people for more. "You set 20% of your packets last month to have QoS high priority, while your plan only allows 10%". While worms will have to do more work to infect hosts, they'll just be a lot smarter about checking hosts. And the home user, even if they got 1:1 IP mappings, will probably stick a nice firewall in front of their modem that blocks incoming packets. Cablemodems (not sure about ADSL) can also be blocked from recognizing more than N MAC addresses per boot, too, so you'll have to alias your NIC to have more IPs (how many home users can do THAT? And it makes routing so much more fun!).

Nothing will change, really, other than not being able to run out of IP addresses. Business as usual.

Hell, NAT has had one benefit - it's made firewalls a lot easier to configure because you don't have to open 20 ports to play a game like you used to just over a decade ago. Torrent clients seem to work fine using one port rather than one port per torrent like they used to. Online gaming seems to work just fine with 2 or 3 ports opened (or none - it was ironically easier to configure my PS3, Xbox360 and Wii to play online than my PC - and I have UPnP disabled!), and many protocols that required incoming connectivity got phased out or adapted (e.g. FTP). And the prevalence of ssh makes life a lot easier for remote access and poor-man's VPN stuff.

Re:I still don't like IPv6

[Locklin]

XMPP (jabber), SIP, Skype, Google Wave, Opera's Unite.

There are plenty of indications of how this can work. Every one of these, however, currently uses horrible, inefficient, and often inconvenient and expensive hacks to send messages between computers.

Re:I still don't like IPv6

[glindsey]

What it's supposed to mean is that every computer can have a public address. So if you sign up with one of the dynamic DNS providers (which will probably be integrated with your OS fairly soon) you should be able to share pictures and things from your own computer without having to upload them to somewhere, or be able to log in remotely to look at some file (private) you forgot to bring with you, or any number of other things (fewer firewall errors on p2p networks? true p2p voip, without needing to sign up with a service that lets you punch holes in NAT?). This would also work without the dynamic DNS provider, but the URL would look uglier.

Most likely, this would also lead to relaxing the typical rule ISPs tend to have against running servers on home connections. They can't really forbid something that gets built into the OS like these sorts of features probably will.

Sure they can! There is absolutely no way Comcast will give users more control over what they can do with their Internet connection. All the big ISPs have tried for years to exert more and more control over what we can do with our connections -- first via social means, then via technical ones. Get everybody using Webmail. Have everybody post videos on Youtube and photos on Flickr. Use pre-made blogging sites, don't host your own.

I'm not saying these are bad sites -- hell, I use them regularly. But they all lead to what the ISPs would love: let people communicate via port 80, period, no listening whatsoever.

I predict IPv6 will lead to Comcast forbidding routers on their network, charging for each additional IPv6 address they hand out, and blocking all listening ports.

But if I don't like it, I can always find another ISP... oh wait, I can't, because they're the only game in town, and they've purchased laws to make sure they stay that way.

Re:I still don't like IPv6

[DragonWriter]

So instead of going to flickr I have to know and maintain all of my friends' computer addresses? In what, an address book that I store on my computer? What if I'm at a friend's house and want to show them another friend's picture, but they don't know the address?

If you were going to do this, yeah, you'd store the address book on your computer (ideally, in the form of a web page or something similarly easy for remote access), and if you were at some other computer and needed to get to one of your favorite sites, you wouldn't need to remember the site, only your own address, and you'd go there, get the link, and go to the site you are looking for.

Note, I actually think the idea that every computer will get used as a server like this with IPv6 to be ridiculous (the reason you don't want every computer used this way is because you don't want your desktop/laptop on all the time, drawing power constantly and busy serving remote requests when you want to use for something else, not because there aren't enough IPs -- it makes more sense for a home to get one, low power, always-on server on the home network for anything that it makes sense to share within that network and/or expose (securely) remotely, but continue to share stuff publicly through external servers, even with IPv6.)

Re:I still don't like IPv6

[Kadin2048]

> I would prefer an addressing system that simplifies life for me.

IPv6 probably will, by virtue -- ironically -- of having more complex addresses. Since you can't reasonably expect people to remember a 128-bit value, even when expressed in hex (hell even if you used Base-64), there will be a far greater reliance on automatic addressing schemes under IPv6 than under IPv4. That's a good thing; Appleshare had the right idea in 1984, and the adoption of IP was a step backwards in this regard (although IP is superior in just about every other way, don't get me wrong, and Appleshare wouldn't have scaled like IP). The length of the IP address shouldn't matter, because the user should never encounter them anyway. With IPv4, doing this automatically was optional because manual addressing was, although inconvenient, admittedly practical. With IPv6 it won't be except in real edge cases, and we'll finally be able to start ignoring IP addresses the same way we basically ignore Ethernet MACs. (Quick! What's your Ethernet MAC? You don't know, more than likely, and it doesn't matter at all. That's how IP should be as well.)

And unlike IPv4, those addresses will be end-to-end routable, which means a lot less fussing around with port forwarding. Plus, the huge number of addresses ought to get rid of dynamic addressing completely, at least the sort of addresses that change themselves automatically from time to time when you disconnect and reconnect. (The addresses will still be "dynamic" in the sense that they're comprised of some part handed down from upstream and some part that's specific to the device, but they won't need to be doled out from some small pool with short leases.) That means no more dyndns.org and abnormally short TTLs to get a usable DNS.

VoIP will be far simpler as well. VoIP over IPv4 is a total pain in the ass, and there are a lot of ugly hacks (STUN, I'm looking at you) in order to try and make it work. In many cases, consumer VoIP equipment tries to avoid the whole mess by forcing you to attach it directly to your WAN uplink, ahead of your router -- which can often be nowhere near where you want it, either in terms of physical location or network topology. The hope of VoIP that comes closer to "just working" in the same way that POTS does would be enough to sell me on IPv6, long addresses be damned.

Re:I still don't like IPv6

[Kadin2048]

So instead of going to flickr I have to know and maintain all of my friends' computer addresses? In what, an address book that I store on my computer? What if I'm at a friend's house and want to show them another friend's picture, but they don't know the address?

That's no harder than knowing his username on Flickr, or the URL of his photostream. You already need to know some piece of information to be able to find his photos on the web; it doesn't matter (from the perspective of the person viewing the photos) whether that URL points to Yahoo's servers or one in his house. You'll probably get it via email or IM and click on it either way.

Re:I still don't like IPv6

[Lumpy]

Funny, I have all that now.

I dont need to buy a service to punch holes in my NAT. I can do that easily and dyndns.org works great.

all 100% free and works if you have at least a slight education in computers. If you dont even have an education, get the HP stuff-server and it does it all for you.

Why the hell do I need ipV6 and a set of live IP addresses on the internet? I need only 1, and I vpn in to keep all the damned punks out of my crap.

Re:I still don't like IPv6

[Kadin2048]

The decision to use hex quad notation to represent IPv6 isn't really that important from a technical perspective. It's important in the short run that IPv6 addresses look different from IPv4s, to distinguish one from the other. However, in the long run it's quite possible to use an IPv4-type notation scheme with v6.

IPv6 addresses will in most instances consist of two components, one which will typically come down from an ISP (the "global prefix" is the proper name, IIRC) and consists of 24 bits, followed by a subnet identifier (8 bits), and then an interface identifier (32 bits).

In most cases where a user is calling their ISP for help, the global prefix will already be known by the tech on the line. It's not necessary for the user to report it. So, if IPv6 addresses become an unbearable burden (which I really don't think they are; people are well-trained to read seemingly meaningless numbers over the phone to tech support), you could quite easily adjust user-facing software to display the subnet and interface identifier in a friendly, IPv4-style form. In this form, the subnet+interface address would look just like an IPv4 address with one extra group of octal digits.

Frankly, I don't really anticipate many reasons for users to need to communicate their IP addresses to tech support once IPv6 becomes the norm. You used to get that a lot because automatic address assignment was a bolt-on to IPv4 and often didn't work right; with IPv6 it's the idea from the very beginning. Plus, IPv6 does a much better job of handling multicast and broadcast traffic, which means you can have automatic service discovery and automatic machine discovery without broadcast storms and all the other issues that have plagued attempts at such systems on IPv4.

Re:I still don't like IPv6

[paul248]

The majority of people have never typed an IP address. Computers use them quadrillions of times a day. I don't think we should cripple the scalability of our communication system just so it's slightly easier for people to look at its inner workings.

In any case, memorizing common IPv6 addresses isn't that hard. [2001:4860:b006::68], for example.

Re:I still don't like IPv6

[MaerD]

...Have you ever spent time on the phones doing tech support?

I've had conversations where relaying the command "cd" took 15 minutes. "ok I need you to cd /etc, that's c as in charlie" "b as in boy?" "no, c as in charlie" "b as in boy?" "no, c as in cookie".

After a few go arounds of "c as in cookie" my cubemate started going "is good enough for me".

now "the tech will know the prefix" doesn't help. What are you going to do? "I only need the last two sets of digits between the colons?" It won't help. People don't deal well with "the computer is down" in the first place, much less when they have to read out crap that doesn't make sense. Numbers are easy, numbers mixed with letters... you're just going to further alienate users and have to hide it deeper to make it "friendly".

Re:I still don't like IPv6

[Anonymous Coward]

If you didn't have an overzealous firewall in the first place, there woulld never have been a problem. This obsession with that kind of firewall seems to date from when Windows left ports open all over the place that were easily exploitable, but that seems to have stopped now. I can't see any reason to fireall off ports > 1023 at the router - I'd just run some sort of local firewall to make sure programs don't open ports unless I want them to.

Without workarounds, NAT constrains things to an "initiate outgoing connection only" model. Look at all the hoops that Skype jumps through to get around it, including the ghastly "supernode" behaviour. And workaround like UPnP defeat the supposed security advantages (I know you said you don't use it, but most people do, myself included).

Finally, and most importantly, we've got NAT over the place already and IP addresses are still on course to run out in 3 years or so.

Re:I still don't like IPv6

[The Moof]

255(255)255-2555

Who would ever put numbers before an area code? Sounds like an idea foreigners would come up with!

On a more serious note, they (the phone company) already do this type of thing. When I was growing up, it was 7-digit dialing, but the number capacity reached and they had to overlay area codes, forcing us all into 10-digit dialing. There was a little bit of backlash at the time, but nobody really cares now.

Re:I still don't like IPv6

[Cajal]

Sure IPv6 has it all. But I doubt any ISP will do business any differently with IPv6 than otherwise. In fact, they'll just salivate that any caps will be reached a bit quicker because of the increased IPv6 header size. Mobile operators are probably salivating as well - 5 cents per kilobyte (not kiB), which includes the OTA headers, plus increased IPv6 header size

Could we please stop it with the baseless assertions that the extra 20 bytes in an IPv6 header will cause so many problems? There is no evidence at all to support this claim. The best example of now IPv6 header size is a non-issue is a paper about adding IPv6 support to OpenMPI. The developers investigated potential performance impacts of running MPI over IPv6 in a cluster. They found a whopping 1.4% drop in throughput and no increase in latency (section 4.1). If IPv6 shows such little hit in a such a sensitive environment, I'm very confident that it will work fine in broadband and cellular access networks.

Further, if you were really so concerned with bandwidth, you'd stop using HTML, since it's a remarkably inefficient encoding system.

Re:I still don't like IPv6

[sjames]

Trying to get "what's your ip address?" when doing telephone technical support is going to be nightmarish.

So you're saying it'll be no worse than now? My IP address is 127.0.0.1.

No, that's the wrong one. Do you see another one?

OH, you're so right, it's 255.255.255.0

Let's try again, that was your netmask.

Let's see, there's 10.1.1.255...

Re:I still don't like IPv6

[socsoc]

While I agree with you, any user running an os with a /etc directory should hopefully know how to navigate the file structure...

Re:I still don't like IPv6

[Anonymous Coward]

What it's supposed to mean is that every computer can have a public address. So if you sign up with one of the dynamic DNS providers (which will probably be integrated with your OS fairly soon) anyone should be able to share your pictures and things from your computer without having to ask your permission, or be able to log in remotely to look at some file (private) you forgot to password-protect, or any number of other things (fewer firewall errors on p2p networks? true p2p voip, without needing to sign up with a service that lets you punch holes in NAT?). This would also work without the dynamic DNS provider, but the URL would look uglier.

Most likely, this would also lead to relaxing the typical rule ISPs tend to have against running servers on home connections. They can't really forbid something that gets built into the OS like these sorts of features probably will.

There, fixed that for you.

A publicly-accessible IP address is a huge responsibility.
I firmly believe that, given the choice, most consumers would prefer the one-way-window to the Internet that NAT provides over having to fend off all the cyber-criminals of the world from trying to peer inside their home.

Re:I still don't like IPv6

[MaerD]

Unless they fired the admin who knows anything, or are half-deaf and trying it out for the first time, or never use the command line but now have to because the X-windows, she no start.

I cut my teeth at Red Hat, all of the above happened to me at one point or another.

Re:I still don't like IPv6

[ion.simon.c]

(Quick! What's your Ethernet MAC? You don't know, more than likely, and it doesn't matter at all. That's how IP should be as well.)

Today finds me without mod points. It is truely a sad day. :(

REPENT!!

[ObsessiveMathsFreak]

Bbrrrriiiing. Bbrrrriiiing.

You: Hello?

Dependant Relative: My internet isn't working!

You: Is the modem turned on?

Dependant Relative: Yes it IS!! It even says I'm connected with eye-pee-vee-six now. But now none of my programs work!! The man from Comcast said it was an upgrade from than eye-pee-vee-four. I thought six was better than four!? Is it because I'm using Windows 7? Do I need to get Windows 6? And my internet is explorer 8? Can I still get emails? And the computer is really slow! Can you come over? ... etc. etc.

You: Curse you Comcast. Curse you!!!

Re:REPENT!!

[sakdoctor]

Wow. I read that as redundant relative both times.

Re:REPENT!!

[thesandtiger]

My brother in law used to call me up, frequently, to ask me for tech support help. He's a doctor, so I solved it by calling him up every single day to ask him some inane question about medicine.

"Hey, so I'm at the store and I want to buy band-aids. Which ones are best?"
"Hey, it's me again - so when I called up 5 minutes ago to ask about band-aids, I didn't realize they had purple ones. Are those going to work differently than the beige ones?"
"Oh, hi, me again... I was walking by the frozen food section and it was kind of cold there but it's a really hot day outside - can I catch sick from the temperature differential?"
"Yeah, it's... well, this is a bit strange. But I was at work today and one of my co-workers kind of has a limp. Can you tell me what that's from? I don't wanna ask him - let me put him on with you, maybe you can fix him..."
"So I was on a date last night and we went to a used bookstore and I started sneezing. Is that the swine flu? Well, yeah, it was dusty in there, but Oprah was talking about the Swine Flu, and I had bacon the other day so maybe I'm going to ... hello? Helloooo?"

For people who don't have a particular profession, calling them up at odd hours to ask them for tiny favors also works. My next-door neighbor used to ask me for tech support all the time, so I started asking him to pick things up at the store for me, give me rides, loan me odd random items ("Can I borrow one of your bookends?" "Do you have a shoehorn I can use for a couple of days? Mine's in the shop.")

Re:REPENT!!

[garcia]

If I had the ability, I would rate this "+5 You Owe Me A Dry Keyboard"

Re:REPENT!!

[ringdangdu]

"Do you have a shoehorn I can use for a couple of days? Mine's in the shop.")

Excellent line!

Re:REPENT!!

[AndrewNeo]

But I only have OS 10.5, do I have to upgrade to 10.6 to use this IPv6?! Protip: Stupid users are still stupid users.

Re:REPENT!!

[bsane]

Heres the thing.... I'm a mac user, use linux and osx at work, and haven't touched windows in years- and even when I did it was just a company pc for email and such.

I converted a few relatives to Macs, before I realized... these people don't understand computers of any kind. They'll always hassle 24/7 no matter what. They may have fewer questions if they have a Mac, but if they stick to windows I can honestly tell them I have no idea and they should try calling Dell or MS for answers. If I'm the one that convinced them to switch- I'm on the hook for the rest of eternity when they go into walmart and buy the crapware of the day and it doesn't work.

Re:REPENT!!

[Anonymous Coward]

All you got to do is tell them that stuff that doesn't work will crash the computer.

Are we serious this time?

[CobaltTiger]

I've been hearing that IPv4 addresses are "almost gone" for maybe 10 years now.

Re:Are we serious this time?

[MobyDisk]

Fortunately, we have been conserving them and switching to NAT so the problem has lessened. The industry isn't crying wolf. Also, if you live in the US, then you have less of a problem than in a developing nation who didn't get a great big block allocated to them.

But if you want your cell phone, computer, XBOX, and refrigerator to have a unique IP address, then this is necessary. Of course, you probably DON'T want that, but well... that's another discussion. :-)

Re:Are we serious this time?

[ColdWetDog]

I've been hearing that IPv4 addresses are "almost gone" for maybe 10 years now.

It's an Illuminati conspiracy tied into fusion research (and holographic storage). Just watch the obituaries. You'll eventually see the pattern. By then it will be too late - another 10 years.

(I'm sure I read it somewhere around here).

Re:Are we serious this time?

[at_slashdot]

I've been hearing that IPv4 addresses are "almost gone" for maybe 10 years now.

That's why we will be unprepared when it finally happens.

Re:Are we serious this time?

[QuantumRiff]

There is alot more to IPv6 then just its IP Address space. there is lots of improvements to security, configuration, and multicasting. Also, the way it is designed will take a HUGE load off the core routers, and actually make them faster... Right now the address space is so fragmented, there are huge tables in them to parse on what subnets are down which paths...

Re:Are we serious this time?

[Macrat]

Don't confuse the mindless tech bashing with actual FACTS!!!

Re:Are we serious this time?

[AndrewNeo]

Sadly enough this is very true. There's no good reason routers (Linksys, Netgear, etc.) shouldn't have at least disabled IPv6 support, but they don't (at least from what I've seen)

Re:Are we serious this time?

[ColdWetDog]

Informative? Oh my.

Note to mods: Please turn off your television sets. They aren't helping you at all....

Re:Are we serious this time?

[thegameiam]

As someone who regularly configures and manages large networks in IPv4 and IPv6, I'd love to hear exactly which improvements you mean. Perhaps you mean the need to have both DHCPv6 AND RA for SLAAC running all the time on a LAN (or you have to cheat off of IPv4 for DNS server addresses); perhaps you mean the lack of RA filtering capability, which makes m-i-t-m attacks a heck of a lot easier; perhaps you're not familiar with v4 multicast.

As for the fragmentation of IPv4 address space, yes, that's a concern if you are running routers which operate in the DFZ. If you aren't, then it's not such a big issue. In any case, routing table growth is likely to be substantially worse in the v6 space than in the v4 space, unless the regional registries can work out policies to keep the number of routes announced to the DFZ to be on the same order of magnitude as the number of ASNs assigned.

In short, there's lots wrong with v6 - the ONLY absolute plus for it is the larger address space. All the rest is consultant-foo.

Re:Are we serious this time?

[j h woodyatt]

Yes. We are serious this time.

Re:Are we serious this time?

[sjames]

You may be reading alarmist press. The more level headed have been saying they will run out one day. Then the word was they won't last long into the next decade. A year or two the estimate was 2012. This year the estimate is still 2012.

Just because you can see the train coming from 10 miles away doesn't mean you should wait until the engineer is frantically sounding the horn and locking up the brakes to step off of the tracks and it sure doesn't mean that when he does he's just being an alarmist. (yeah, yeah, there's been a train coming and we haven't been hit YET!)

Re:Are we serious this time?

[SpazmodeusG]

Where do you live?
If you live in America then please shut the fuck up. Those of use in the rest of the world are starting to be allocated private addresses that are being NATed via our ISPs. This means we can't do port forwarding or use our connections for anything other than browsing on port 80.

http://forums.whirlpool.net.au/forum-replies-archive.cfm/981410.html

What?

[geekoid]

Verizon has IP6.

Re:What?

[Randle_Revar]

For residential users?

Nobodies asked yet

[Anonymous Coward]

Just how much extra are they going to charge the customers for the privalege of of a cost of them staying in business.

services?

[Neil Watson]

Potentially these customers will have a small block of ipv6 addresses. Will they be allowed to run their own web or email services?

Re:services?

[characterZer0]

No. They'll get one address. And they still will not be able to run services.

What's the big deal with IPv6

[pak9rabid]

Why does everyone here get so excited when anything about IPv6 is mentioned? From an end-user's perspective, it appears to accomplish the same thing that IPv4 does, except addresses are longer and contain more characters. Are there any real benefits from and end-user's perspective in using IPv6? ISPs are still going to charge the same amount for public IPs and people are still going to user routers with NAT to save money on having to pay extra for additional IPs. From a sysadmin point of view, it's just going to mean more work and probably sleepless nights as we discover quirks with software and equipment that don't play nicely with IPv6. So, whats to get excited about?

Re:What's the big deal with IPv6

[Anonymous Coward]

ISPs are still going to charge the same amount for public IPs and people are still going to user routers with NAT to save money on having to pay extra for additional IPs.

That would be quite pointless, given the number of IPs available. Why shouldn't the ISP just hand out a /64? There are plenty of them to go around. The ISPs gave up on the idea of trying to make extra money from multiple devices connected a while ago - and since they know people will just use NAT if they only give out one IP, why bother?

Re:What's the big deal with IPv6

[Attila Dimedici]

ISPs are still going to charge the same amount for public IPs and people are still going to user routers with NAT to save money on having to pay extra for additional IPs.

That would be quite pointless, given the number of IPs available. Why shouldn't the ISP just hand out a /64? There are plenty of them to go around. The ISPs gave up on the idea of trying to make extra money from multiple devices connected a while ago - and since they know people will just use NAT if they only give out one IP, why bother?

Because there are people who NEED that additional public IT. Those people will pay to get it, since NAT doesn't work for whatever it is they are doing. Additionally, there are a lot of people whose LAN would be screwed up by having all of their machines have a public IP address and who don't know enough to fix it.

Re:What's the big deal with IPv6

[MadKeithV]

Because 640k is enough for anybody!

Re:What's the big deal with IPv6

[pak9rabid]

That would be quite pointless, given the number of IPs available. Why shouldn't the ISP just hand out a /64? There are plenty of them to go around.

Because ISPs exist to make money, not to provide a civil service to people. ISPs (especially the bigger ones) are going to do whatever they can to maximize profits. Just because there's essentially an unlimited number of IPv6 addresses available doesn't mean that the value of a public IP will disappear.

Re:What's the big deal with IPv6

[quazee]

Yes, in fact, stateless autoconfiguration implies using at least a /80 prefix.
And I don't see why ISPs would want needless complexity of keeping track of every device in a household.

Re:What's the big deal with IPv6

[Anonymous Coward]

I thought it was /64, but either way, it's a tiny fraction of the address space. An ISP with a /32 allocation could give billions of customers a /64 each.

ISPs *could* try giving out private space now, on the grounds that there is a shortage and it's good enough for a lot of users, and sell a genuine public IP as a premium option, yet they haven't done.

Re:What's the big deal with IPv6

[z4ce]

Let's say you're using Skype or bittorrent. And you want to do it on more than one computer, and you want to do it relatively efficiently. You need IPV6. Creating P2P apps is a pain with all of the NAT in the world.

Re:What's the big deal with IPv6

[Just Some Guy]

Additionally, there are a lot of people whose LAN would be screwed up by having all of their machines have a public IP address and who don't know enough to fix it.

No, there aren't. A DSL or cable modem with a default-deny firewall (which will be all of them) will give a superset of the protections NAT offers now. There's a difference between public and publicly routable, you know.

Re:What's the big deal with IPv6

[Dragonslicer]

That would be quite pointless, given the number of IPs available. Why shouldn't the ISP just hand out a /64? There are plenty of them to go around.

Because ISPs exist to make money, not to provide a civil service to people. ISPs (especially the bigger ones) are going to do whatever they can to maximize profits. Just because there's essentially an unlimited number of IPv6 addresses available doesn't mean that the value of a public IP will disappear.

Doing something like that might get (and would most likely deserve) an investigation by the FTC and/or federal Department of Justice. If every major ISP charged extra money for a resource that has no practical limit, you'd have a fairly easy collusion and price-fixing case. If you can have such a case against memory manufacturers, who deal with creating physical items that are obviously much more limited than IPv6 addresses and actually require money and resources to create, even a Slashdot I-am-not-a-lawyer could win a case over something like IPv6 addresses.

Re:What's the big deal with IPv6

[AndrewNeo]

This. The point of assigning a /64 is to prevent the need to NAT.

Re:What's the big deal with IPv6

[Hatta]

Why does everyone here get so excited when anything about IPv6 is mentioned?

Two words: No NAT.

Re:What's the big deal with IPv6

[kimvette]

A lot of devices still do not support IPv6. Phones, cellphones,

A lot of people have to type in IP addresses (sysadmins, etc.) when configuring devices, DNS, web servers, and so forth, and those huge address strings are a pain in the ass. I don't want to deal with them. I like the dotted quads.

Also, one occasionally needs to access machines by IP address when DNS flakes out. What do you do when a DNS server goes down? Ideally you have a secondary DNS however not all organizations are willing to spend the money - especially in this economic climate.

Re:What's the big deal with IPv6

[amorsen]

Because ISPs exist to make money, not to provide a civil service to people. ISPs (especially the bigger ones) are going to do whatever they can to maximize profits. Just because there's essentially an unlimited number of IPv6 addresses available doesn't mean that the value of a public IP will disappear.

It all depends on how much people are going to scream. Customers of large corporations generally get screwed precisely as much as they accept without making public protests.

Hopefully people will complain loudly enough that the ISPs will just follow the path of least resistance (i.e. the standards), since it doesn't cost them anything.

Re:What's the big deal with IPv6

[Anonymous Coward]

http://3232236033/
Login: admin
Pass: *****
click. virtual servers
6881:192.168.2.47:6881
6882:192.168.2.42:6882

Of course, it would be better to just put your torrents all on one "server" machine and keep an instance of btlaunchmanycurses watching a directory on it (add or remove stuff from the directory at will, of course). That way it can manage your connection and make sure every torrent gets a chance.

Of course you're only using that to distribute your collection of pictures from little jimmy's birthday party, linux isos, and other things you have legitimate copy rights over, right?

Re:What's the big deal with IPv6

[Anonymous Coward]

"ISPs are still going to charge the same amount for public IPs and people are still going to user routers with NAT to save money on having to pay extra for additional IPs."

I'm excited because the above statement is false. Now please excuse me, I'm off to the IPv6 parade.

Re:What's the big deal with IPv6

[Anonymous Coward]

Why does everyone here get so excited when anything about $NEW_TECHNOLOGY is mentioned? From an end-user's perspective, it appears to accomplish the same thing as $OLD_TECHNOLOGY, except {faster|smaller|cheaper}. Are there any real benefits from an end-user's perspective in using $NEW_TECHNOLOGY? etc etc. So, whats to get excited about?

I don't think you quite fit in here.

Re:What's the big deal with IPv6

[sjames]

If they hand out a /64, your computers will autoconfigure. If not, they get to talk many many thousands of people through manually configuring an IPv6 address.

That costs real money.

Re:What's the big deal with IPv6

[knorthern knight]

> That would be quite pointless, given the number of
> IPs available. Why shouldn't the ISP just hand out
> a /64? There are plenty of them to go around.

I'm sure TPTB (The Powers That Be) were thinking exactly the same thing 35 years ago as they handed out class A and B blocks of IPV4 addresses to companies and universities.

It's possible for Comcast to run their entire US operation with on /48. Shit happens. We might save ourselves a lot of headaches several years down the road if we treat IPV6 addresses as scarce commodities from day 1.

Re:What's the big deal with IPv6

[hab136]

Let's say you're using Skype or bittorrent. And you want to do it on more than one computer, and you want to do it relatively efficiently. You need IPV6. Creating P2P apps is a pain with all of the NAT in the world.

Not exactly the greatest of examples. Skype is the king of NAT busting, no user intervention required.

Torrents need one port forwarded per machine at the router, which is either not that hard, or automatic. Many torrent clients can use uPnP or NAT-PMP to map the port automatically.

Re:What's the big deal with IPv6

[z4ce]

Skype tries to NAT bust. But it still works MUCH faster without it. Also, many, many home routers choke when keeping up with bittorrent connections. The point is two computers are behind NAT there has be a third-party (non-NAT) involved to negotiate for us. And even then, depending on the "routers" involved it might not work for creating a direct connection.

OMG! OMG!.IPv6 is coming for ME!

[ae1294]

It's funny how all of you are complaining so much about this. IPv6 is a required evil for the internet to keep going and it will simplify things greatly and should speed up things in general too. That is if and when they get rid of the IPv4 hardware...

I've never seen a bunch of self described computer geeks whining so much about something that will simplify routing and get rid of NAT which is a truely horrid hack.

Come on guys, you know you are going to have to deal with problems no mater what happens in computer land?! Might as well deal with a problem that will make the internet routing make sense again and it's not like it will need to be done again in your life time.

Re:OMG! OMG!.IPv6 is coming for ME!

[DigitalSorceress]

NAT and PAT may be "hackish" but I for one am really glad that they have gotten such widespread use. For the vast majority of non-techie Internet users, a simple D-Link, Linksys, etc... firewall/router with its fairly transparent PATing is a nice bit of security that they have even if they don't understand it.

Also, the ISPs used to be really weird about home networks, but over time, they've changed their attitudes. If they fully implemented IPv6 to the point where every device could have its own publicly routable address, I'd be really worried about the security, PLUS I bet you that Comcast and the like would start CHARGING BY THE IP.

I'll stick to my router and its Port Address Translation even if they switch over to IPv6.

Us self-described techie types aren't afraid of change per-se, it's just that an awful lot of us have been burned by living on the bleeding edge, so when it comes to something that's important (to us) to keep up and running in a stable manner, we tend to fall back to "tried and true".

I bet us IPv6 refuseniks who also happen to be Windows users / admins are also the types to NEVER trust an MS service pack until we've tried it out on one guina pig machine and until its been out in public long enough to see if it's a shitstorm in a service pack's clothing.

 

Re:OMG! OMG!.IPv6 is coming for ME!

[Monkeedude1212]

LoL. I'm not complaining about IPv6 because I know its not going to get rid of NAT tables. IPv6 has somewhere around 270 quadrillion available address combinations (255^6 right? subtract a couple for those obvious reasons)

Nothing doesn't make sense with NAT. In all honesty its how the whole thing should have been structured in the first place.

Why?

If one day we reach more then 270 quadrillion networked computers (that'd make for an awesome lan party, but I shouldn't get distracted) we'll just have to shift over to IPv8. Doing that on alot of PC's won't be as smooth as this IPv4 to IPv6 will be.

But guess what, if you understand NAT, you will NEVER have to upgrade past IPv4, because you will NEVER run out of IP Addresses. NAT is just the flexible approach to the problem that alot of people don't like because they don't understand.

Re:OMG! OMG!.IPv6 is coming for ME!

[Just Some Guy]

But guess what, if you understand NAT, you will NEVER have to upgrade past IPv4, because you will NEVER run out of IP Addresses. NAT is just the flexible approach to the problem that alot of people don't like because they don't understand.

Meanwhile, back in reality...

In abstract, NAT treats addr+port as a 48-bit address, so you're effectively trading ports for address. That means you only get one port 80 per public IP, so forget having more than one webserver (unless you can somehow get your visitors to go to http://www.example.com:8080/ ). Every P2P app, every Skype, every game server, every random application you want to post has to have a unique port number across your entire network.

Can you really not see why that sucks in comparison to IPv6 which lets every machine on your LAN listen on the whole 2^16 port range as your firewall allows?

People who don't understand NAT at all like IPv6. People who only barely understand it, like yourself, think IPv4+NAT is spiffy. People who actually understand NAT and what it implies think that it needs to be taken out back and shot.

Re:OMG! OMG!.IPv6 is coming for ME!

[asdfghjklqwertyuiop]

(255^6 right? subtract a couple for those obvious reasons)

No. 2^128. Indescribably more. It seems you are the one who doesn't understand a few things.

Re:OMG! OMG!.IPv6 is coming for ME!

[asdfghjklqwertyuiop]

For the vast majority of non-techie Internet users, a simple D-Link, Linksys, etc... firewall/router with its fairly transparent PATing is a nice bit of security that they have even if they don't understand it.

Those things can provide just as good of security some time even without complexity of NAT. Try setting up a machine outside your NATing router/firewall and add a route to the private network via that firewall/router. Notice how the packets still don't make it in even though you've just negated the obscurity of NAT...

Re:OMG! OMG!.IPv6 is coming for ME!

[BitZtream]

I think its funny how you think you know what you're talking about.

First off, people use PAT, not NAT, very VERY few people use NAT. NAT maps an IP address on one side to an IP address on the otherside. Each internal address uses an external address exclusively.

No ISP does this for typical home subscribers, you get one IP so NAT would fail horribly.

What you actually use is PAT, Port and Address Translation. This is the 1 to many that you see people using so they can get a bunch of machines on the Internet with only a single external address. Yes I realize its common to refer to it as NAT, but thats wrong and so are you, and if you're going to act like an arrogant pickle smoker I'm going to have to call you on it.

Second, IPv6 will do nothing to make routing easier, the statement in and of itself is retarded and shows a complete lack of practical understanding of why routing is the mess it is.

When the Internet started (pick a point in time, any point doesn't really matter when at all), they didn't sit down and assign all the IP addresses to everyone and lay out the entire mesh of connections around the world to facilitate the most efficient topology did they? No, the didn't, it evolved over years. Connections come and go, address assignments come and go or get reassigned, thats just the way of the Internet, it changes.

IPv6 in no way deals with that fact. It gives us enough addresses assign 2^52 of them to every visible star in the sky and have plenty left over. It deals with the fact that we're low on addresses by giving us a lot more. It gives us some ways to auto-assign addresses and blocks based on where they are connected. What it does not deal with is that you are in Dallas, and your connections for auto-assignment are in new york and san francisco. It doesn't deal with what happens when you move that subnet to new york.

IPv6 will do nothing at all to simplify routing. There is more to the routing problem than address assignments and the only people that think it can be fixed by adding a lot more numbers and subnets to the mix are just illustrating their complete lack of grasp of the actual problem.

Please do not try to educate people about IPv6 or Internet routing/protocols in the future, you're missing some fairly key knowledge that just means you're making it so other people are wrong as well.

Re:OMG! OMG!.IPv6 is coming for ME!

[digitalsushi]

Your last sentence claims that we'll never have to upgrade from v4 since we can use address translation, because it's a flexible approach. Yet these techniques fall apart when a remote destination wishes to establish communication with a translated address. Horrible technologies like STUN enter the picture, requiring a third node with a public address to establish the network parameters for translated hosts. What would occur when we can no longer even provide addresses to these intermediaries? This is to say nothing of the knowledge higher-level protocols must have of those layers beneath, which is unwholesome and breaks the abstract nature that once facilitated development. Metaphorically, your postman has to know what's inside the box in order to deliver it. Is this the flexibility you laud? This is why I don't like it. Am I misinformed?

Re:OMG! OMG!.IPv6 is coming for ME!

[ae1294]

NAT and PAT may be "hackish" but I for one am really glad that they have gotten such widespread use. For the vast majority of non-techie Internet users, a simple D-Link, Linksys, etc... firewall/router with its fairly transparent PATing is a nice bit of security that they have even if they don't understand it.

Yes it does add a tinny amount of security but it wasn't really designed to do that. It was designed to get around the fact that there aren't enough addresses to go around. Maybe we as IT professionals should educate people about the need to secure their systems. Maybe just maybe if people saw their brand new DELL,HP,Compaq,etc computer get taken down 10 seconds after they connect it to a direct internet connection after buying it they will learn why it's really important to install those annoying updates that come out every few weeks. Right now I have family member's who actively try and ignore those little messages that pop up at the bottom of their WinXP boxes. That isn't really a good thing and a firewall doesn't protect you once you start googling for offshore casino's and porn. And I think we all know that most non-geek computers are infected with some kind of spyware so your security device isn't really doing that great of a job.

Also, the ISPs used to be really weird about home networks, but over time, they've changed their attitudes. If they fully implemented IPv6 to the point where every device could have its own publicly routable address, I'd be really worried about the security, PLUS I bet you that Comcast and the like would start CHARGING BY THE IP.

Yes they really where and you are right they will try and do everything they can to make a buck but the extreme amount of addresses provided is going to make it hard to justify a charge. I remember back o 10+ years ago when I had COX Cable and they charged 3 bucks for an IP. Now I can't get an IP address unless I pay the $400 for business service {Charter}. So basically the Firewall/routers you speak of have kind of screwed me because I'd really like to have a couple of real IP's.

I'll stick to my router and its Port Address Translation even if they switch over to IPv6.

PLEASE no you won't... That's like saying you are going to keep your 386sx system because your scared of the new 486 processors and don't want the extra processing power.

Us self-described techie types aren't afraid of change per-se, it's just that an awful lot of us have been burned by living on the bleeding edge, so when it comes to something that's important (to us) to keep up and running in a stable manner, we tend to fall back to "tried and true".

What bleeding edge? IPv6 has been out for 10 years! that isn't bleeding edge man. Maybe you're just worried about having to learn some new-er stuff? Maybe have to ask some questions that make you look foolish? eh?

I bet us IPv6 refuseniks who also happen to be Windows users / admins are also the types to NEVER trust an MS service pack until we've tried it out on one guina pig machine and until its been out in public long enough to see if it's a shitstorm in a service pack's clothing.

Really? cause last time I checked 99% of service packs for windows are just all the security updates that have come out over a year or so. If you are saying that every time Microsoft releases a security patch you break out a Guinea pig machine than I really feel sorry for you. MS Security patches come out weeks or months after the bad guys figure out an exploit so the rest of us wait maybe a week on service pack and no time on patches. Everything works just fine 98% of the time except for maybe a program or two and as we tend to do regular backups if something does go horribly wrong then we restore from our backup! Plus most updates have an uninstall anyhow.

Re:OMG! OMG!.IPv6 is coming for ME!

[ae1294]

I think its funny how you think you know what you're talking about.

Same here?

First off, people use PAT, not NAT, very VERY few people use NAT. NAT maps an IP address on one side to an IP address on the otherside. Each internal address uses an external address exclusively.

please... if the Linux man pages for ipchains and iptables call it NAT then I'm going to call it NAT as I pay my respects to 'the man' regardless of what vocabulary words are on the A+ Cert now days.

Second, IPv6 will do nothing to make routing easier, the statement in and of itself is retarded and shows a complete lack of practical understanding of why routing is the mess it is.

I disagree, firstly while slightly unrelated to what you are talking about a prime example of the problem is P2P programs. Have you tried to download a torrent on a Linksys Router? Yeah try downloading a few at the same time and watch the router thrash because it can't keep up with all the connections and then go monkey with the torrent program options and set your max connections to something pathetic like 500. What's the point of future 100Mbit internet if you can't download anything faster than 10Mbit because your NAT tables get clogged.

facilitate the most efficient topology did they? No, the didn't, it evolved over years. Connections come and go, address assignments come and go or get reassigned, thats just the way of the Internet, it changes.

ummmmm well yeah.. they sorta did.. now days you have bits of every Class A,B and sometimes C networks all over the place. (not that there really is such a thing now days) Well that's all fine and dandy if you are a company and that's how you wanna setup your address space. It's not good for the internet backbone that it has to be done because there are too few addresses.

It deals with the fact that we're low on addresses by giving us a lot more. It gives us some ways to auto-assign addresses and blocks based on where they are connected. What it does not deal with is that you are in Dallas, and your connections for auto-assignment are in new york and san francisco. It doesn't deal with what happens when you move that subnet to new york.

Well ok then what is your idea for correcting this? How does this issue have anything to do with the fact we need more address space?

Please do not try to educate people about IPv6 or Internet routing/protocols in the future, you're missing some fairly key knowledge that just means you're making it so other people are wrong as well.

yes your right.. we should tell everyone that IPv6 is bad and to stay away from it because it's not perfect and computers must be perfect. Maybe we can start charging more for internet access because there aren't enough addresses for everyone to be online at 5pm in CA and thus the internet can be just like GAS is today.... BRILLIANT!

Re:OMG! OMG!.IPv6 is coming for ME!

[kindbud]

It's funny how all of you are complaining so much about this. IPv6 is a required evil for the internet to keep going and it will simplify things greatly and should speed up things in general too. That is if and when they get rid of the IPv4 hardware...

IPv6 was designed specifically so you don't have to get rid of all your IPv4 gear.

I've never seen a bunch of self described computer geeks whining so much about something that will simplify routing and get rid of NAT which is a truely horrid hack.

NAT is an integral part of IPv6. An IPv6 prefix can be large enough to encompass a block of addresses as large as the entire IPv4 address space. So you connect your IPv4 nodes to the IPv6 internet with a NAT box that prepends your IPv6 prefix to the IPv4 address to form a unique IPv6 address, and NAT that through your gateway.

Re:OMG! OMG!.IPv6 is coming for ME!

[srvivn21]

But guess what, if you understand NAT, you will NEVER have to upgrade past IPv4, because you will NEVER run out of IP Addresses. NAT is just the flexible approach to the problem that alot of people don't like because they don't understand.

Meanwhile, back in reality...

In abstract, NAT treats addr+port as a 48-bit address, so you're effectively trading ports for address. That means you only get one port 80 per public IP, so forget having more than one webserver (unless you can somehow get your visitors to go to http://www.example.com:8080/ ).

Incorrect. Had you said you only get one port 443 per public IP, I wouldn't have an issue, but HTTP traffic is easy to "route".

Every P2P app, every Skype, every game server, every random application you want to post has to have a unique port number across your entire network.

Can you really not see why that sucks in comparison to IPv6 which lets every machine on your LAN listen on the whole 2^16 port range as your firewall allows?

No argument from me here.

People who don't understand NAT at all like IPv6. People who only barely understand it, like yourself, think IPv4+NAT is spiffy. People who actually understand NAT and what it implies think that it needs to be taken out back and shot.

Every tool has a purpose. NAT is fine for a home, a small business, or an arbitrarily large network of strictly client computers.

Re:OMG! OMG!.IPv6 is coming for ME!

[ae1294]

IPv6 was designed specifically so you don't have to get rid of all your IPv4 gear.

UUUGGGHHH... I KNOW! ... It was designed so we don't have to shut the internet off for a year to move everything over to IP6....

NAT is an integral part of IPv6. An IPv6 prefix can be large enough to encompass a block of addresses as large as the entire IPv4 address space.

It's not integral if you don't need IP4 anymore but yes like I said IP6 is wonderful and you can roll it out without shutting down the whole inter-web which is why it should already be deployed some 10 years after being created and approved as the next protocol.

Re:OMG! OMG!.IPv6 is coming for ME!

[Just Some Guy]

Incorrect. Had you said you only get one port 443 per public IP, I wouldn't have an issue, but HTTP traffic is easy to "route".

I hearby declare Squid's reverse proxy mode as another form of NAT and wave my hands at it. :-)

Every tool has a purpose. NAT is fine for a home, a small business, or an arbitrarily large network of strictly client computers.

I'd still provisionally disagree. NAT breaks things that homes, small businesses, and even strictly client networks might like to use (such as P2P and VOIP).

Re:OMG! OMG!.IPv6 is coming for ME!

[Anonymous Coward]

That means you only get one port 80 per public IP, so forget having more than one webserver (unless you can somehow get your visitors to go to http://www.example.com:8080/ [example.com] ).

Apache VirtualHost directive?

Re:OMG! OMG!.IPv6 is coming for ME!

[Just Some Guy]

That only handles multiple domains on the same IP, not multiple servers. You have to switch at the HTTP level (as srvivn21 described) to have multiple machines served by the same IP, and then you'd adding a new single point of failure.

Re:OMG! OMG!.IPv6 is coming for ME!

[dwye]

> unless you can somehow get your visitors to go to http://www.example.com:8080/

By publishing your URL as above? You are right. That is far too difficult for anybody, or any program, to use. You would have to save your favorite URLs in a table of bookmarks, or something; perhaps have an HTTP tag to encapsulate other URLs. Both of these ideas are clearly insane.

Re:OMG! OMG!.IPv6 is coming for ME!

[knorthern knight]

>> I'll stick to my router and its Port Address Translation even
>> if they switch over to IPv6.

> PLEASE no you won't... That's like saying you are going to keep
> your 386sx system because your scared of the new 486 processors
> and don't want the extra processing power.

What, exactly, will the benefit of not NATing be to 95% of users? I'm not talking about the bandwidth hogs running bit-torrent or the geeks here running exotic setups. I'm talking about the vast majority that don't run servers.

Yes, NAT doesn't allow servers, except via port-forwarding. And yes it may kill some brain-dead apps. But for most of us it doesn't matter.

> What bleeding edge? IPv6 has been out for 10 years! that isn't
> bleeding edge man. Maybe you're just worried about having to
> learn some new-er stuff? Maybe have to ask some questions that
> make you look foolish? eh?

I run Gentoo linux, as much for the control-freak aspect as for the optimization aspect. One day a few years ago the maintainers, in their "infinite wisdom", decided to make IPV6 a default USE flag. That meant that every app that was capable of IPV6 support got it built. Of course, I noticed it when mplayer, Firefox, etc started taking an extra 45 seconds to start doing their thing, because they sent out IPV6 DNS requests first, and waited for them to time out, before sending out IPV4 DNS requests. Let's just say I was not impressed. From that point onwards, I put in "-*" at the beginning of the USE flag, and followed it with only the items I wanted/needed.

One day, when it makes sense to do so, I will convert to IPV6. That day will be when my ISP starts supporting it. I'll put up a test machine and make sure that my IPV6 iptables firewall is just as anal as my IPV4 iptables firewall. Following that, I'll put a NATing IPV6 router in front of the machine, and migrate over.

Unlike me, most end-users will probably hook up their IPV4 machines with IPV4 software to a converter box. I'm showing my age, but...

- when UHF first came in (channels 14-to-83 back then), people got converter boxes so they wouldn't have to throw their old TV sets away

- when cable "mid-band" and "super-band" channels first came in, people got converter boxes to they wouldn't have to throw their old TV sets away

- when ATSC first came in, people got converter boxes to they wouldn't have to throw their old TV sets away

Notice a pattern? I expect that when IPV6 first comes in, people will get converter boxes to they wouldn't have to throw their old software away.

Re:OMG! OMG!.IPv6 is coming for ME!

[hab136]

That only handles multiple domains on the same IP, not multiple servers. You have to switch at the HTTP level (as srvivn21 described) to have multiple machines served by the same IP, and then you'd adding a new single point of failure.

Two routers (cross connected)
|x|
Two load balancers (HSRP/VRRP to share one IP)
| | | | |
Farm of web load balancers (for example Squid)
| | | | |
Farms of Farm of web servers (5 servers for site A, 7 for site B, 13 for site C)

It can be done, and often is done this way in large enterprises. No single point of failure, only one IP, and web servers need know nothing about the outside world or each other.

You need a new IP for every SSL site, but you can put that on the IP load balancers and map them to different ports on the web load balancers.

Re:OMG! OMG!.IPv6 is coming for ME!

[hab136]

Every P2P app, every Skype, every game server, every random application you want to post has to have a unique port number across your entire network.

So? For the inbound ports, uPnP or NAT-PMP takes the administrative hassle of assigning ports out of the way for applications that don't need a well-known one.

Anyways, not every stream needs a new port number. Consider this example:

10.0.0.1:4444 -> NAT:1111 -> 1.1.1.1:80
10.0.0.1:4445 -> NAT:1111 -> 2.2.2.2:80
10.0.0.2:5555 -> NAT:1111 -> 3.3.3.3:80
10.0.0.3:6666 -> NAT:1111 -> 4.4.4.4:80
10.0.0.1:4446 -> NAT:1112 -> 1.1.1.1:80

There, I just made 5 outbound connections using two NAT ports. The NAT will have to take the remote IP/port into consideration instead of just the port number for its translation, which is more work on the NAT device, but entirely possible.

So now I need a new port for:
1. Every new outbound connection to the same destination IP
2. Every inbound (LISTENING) connection

Let's say you're torrenting 1,000 different Linux ISOs to 4,000 of your closest friends. You could potentially be using 2 ports - one for inbound, one for outbound, since we're only making one outbound connection to each IP. 1 port for 4,000 outbound streams, 1 port for inbound.

Let's add another machine doing the same thing, but with FreeBSD ISOs. If he's sending to 4,000 different people, then you only need one additional port - for the inbound.

1(one!) port for 8,000 outbound streams, 2 ports for inbound.

For web traffic, you'll have a proxy inside your network making all the web requests. Not only will that speed access through caching (and therefore less streams), but the proxy will be able to consolidate streams to the website. Instead of 5 clients opening 5 streams to slashdot.org, the proxy will open one or two and pipeline requests down them.

What usage pattern are you expecting that you'll exceed 65,000 ports behind one IP? Remember these are only inbound listening port or simultaneous outbound streams - the instant that the outbound stream disconnects, the port number is available for reuse.

People who don't understand NAT at all like IPv6. People who only barely understand it, like yourself, think IPv4+NAT is spiffy. People who actually understand NAT and what it implies think that it needs to be taken out back and shot.

NAT is hideous and ugly. It's a pain to program for and administrate. I'd love to see IPv6 replace it.. but NAT does in fact work just fine for now. Learn to live with it, because IPv4 isn't going away any time soon.

Re:OMG! OMG!.IPv6 is coming for ME!

[srvivn21]

Incorrect. Had you said you only get one port 443 per public IP, I wouldn't have an issue, but HTTP traffic is easy to "route".

I hearby declare Squid's reverse proxy mode as another form of NAT and wave my hands at it. :-)

Heh. I suppose it could be called such. In a way.

Every tool has a purpose. NAT is fine for a home, a small business, or an arbitrarily large network of strictly client computers.

I'd still provisionally disagree. NAT breaks things that homes, small businesses, and even strictly client networks might like to use (such as P2P and VOIP).

I figure, once you start a P2P application (or are the recipient of a VOIP call), you are no longer strictly a client. For homes and small businesses, port forwarding is a workable (if inelegant) solution.

Re:OMG! OMG!.IPv6 is coming for ME!

[ae1294]

What, exactly, will the benefit of not NATing be to 95% of users? I'm not talking about the bandwidth hogs running bit-torrent or the geeks here running exotic setups. I'm talking about the vast majority that don't run servers.

Continued use of the internet after address space is exhausted. Using hulu, youtube, myspace music and videos, etc, etc all at the same time on their 4 different computers without having to build a Linux NAT box strong enough to handle the load. You act as like normal people aren't becoming bandwidth hogs... Remember 14.4Kbps back in the BBS days? it was pretty damn fast on those ANSI menus... Have you looked at the size of a normal website lately? Just think about it.. Websites use to be maybe a Meg or two now they are HUGE! Add to that the trend of moving to streaming music and video over the internet to whole families and things only getting larger and larger every day. I will concede that they all could just go out and buy new more powerful linksys routes but hey I personally like to correct problems and IPv4 address space is a problem and rolling it out doesn't have to affect anybody unless we wait until it's too late.

Unlike me, most end-users will probably hook up their IPV4 machines with IPV4 software to a converter box.

They could, or they could also handle it in software with a program raper APP or just a fake network driver that acts as a NAT layer that converts IPv4 requests to v6 and back. I'd rather have bloat code installed on their computer's than on mine...

Re:OMG! OMG!.IPv6 is coming for ME!

[ion.simon.c]

I run Gentoo linux, as much for the control-freak aspect as for the optimization aspect. One day a few years ago the maintainers, in their "infinite wisdom", decided to make IPV6 a default USE flag.

See, that's kinda funny. I've been a Gentoo Linux user since... 2001, 2002 or so. I set up my HE.net tunnel in August of 2008. Prior to that period I had no IPv6 connectivity.
I have not run into the bug that you're describing.

...they sent out IPV6 DNS requests first, and waited for them to time out, before sending out IPV4 DNS requests.

Are you talking about the strict RFC3484 support that was added into Linux back in the early days of 2.6? That could explain the issues that you're reporting. Rest assured in the knowledge that glibc's handling of this has been much more sane for many years now.

Time Warner is already doing this in Brooklyn/NYC

[thesandbender]

However, they're being really evil and routing all their traffic through SWIP's 6 network... Which means everything gets routed over to Amsterdam and then back.  e.g. :

    C:\Users\Mike>tracert -6 ipv6.google.com

    Tracing route to ipv6.l.google.com [2001:4860:b004::68] over a maximum of 30 hops:

    1    <1 ms    <1 ms    <1 ms  2002:185a:90f:1234::1
    2     *        *        *     Request timed out.
    3   109 ms   107 ms   109 ms  ams-core-1.tengige0-0-0-0.swip.net [2a00:800:0:1::1:1]
    4   110 ms   110 ms   109 ms  ams16-core-1.gigabiteth4-0-0.swip.net [2a00:800:0:1::2b:1]
    5   105 ms   109 ms   107 ms  pr61.ams04.net.google.com [2001:7f8:1::a501:5169:1]

Well googles local AMS server handles it but you get the idea.  It's slower and you have to wonder how long before SWIP gets pissed.

Small block?

[XanC]

Unless Comcast is totally bucking well-established standards (which for them is possible, but I really don't see it) then every customer will be allocated a /64. In other words, every customer will have the square of the IPv4 address space to play with.

Seems like they'd have to relax rules on listening ports.

Re:Small block?

[AndrewNeo]

I highly doubt they're going to relax any rules on listening ports (IP allocation has nothing to do with it) but that does make me wonder if they'll be dynamically allocated like they are now or if the IPs will be statically assigned. (Obviously they'll still use DHCP to distribute, but will the IP change like it does now? etc.)

Re:Small block?

[XanC]

DHCP is not necessary for IPv6. Most likely, they'll assign a /64 to your router, and from there, your individual machines will self-assign permanent addresses based on their MACs.

But it's possible that the /64 could vary, I suppose. Hopefully we'll find out soon.

Re:Small block?

[Cyberax]

They'll probably use stateless autoconfiguration (since it's easier). That way Comcast will assign your computer /80 prefix and the rest 48 bits are derived from your MAC address. Or you can configure them manually for a nice-looking number.

Re:Small block?

[amorsen]

Actually, unless Comcast is totally bucking well-established standards, every customer will be allocated at least a /56. Giving out just a /64 is severely frowned upon, and lots of us are crossing our fingers that it doesn't happen.

Re:Small block?

[paul248]

Everything I've heard points to Comcast assigning every house a /56 (that's 256 /64s).

One /64 would be very limiting, because you'd only be able to have one subnet.

Re:Small block?

[Anonymous Coward]

OK, but how many residential users have more than one subnet?

Anyway, you *can* subnet a /64, it just means you can't use stateless autoconfiguration.

Re:Small block?

[j h woodyatt]

They'll almost certainly be using DHCP6 with PD to hand out prefixes in the trials-- probably longer than /48 but shorter than /64. I'm guessing they will probably go with /56 if and when they roll out the commercial service, but I won't be shocked if we end up seeing /60 prefixes instead.

Re:Small block?

[j h woodyatt]

Anybody using an Apple wireless access point might conceivably need two subnets, i.e. one for the private home network and the other for the guest network.

Re:Small block?

[ion.simon.c]

What's wrong with just a /64?

Re:Small block?

[amorsen]

You can't subnet it. Well you can, if you assign everything statically and you avoid certain implementations of IPv6, but it's a bad idea if you want things to "just work" -- and one of the advantages of IPv6 is the way that things "just work".

So with just a /64 you can't make a separate network for guest Wifi, or a DMZ, or for separate departments. All that is easy to do with IPv4 and NAT, so if ISP's only give out a /64, NAT will quickly get popular in IPv6 too.

Believe it when I see it.

[maskedbishounen]

Much like the mythical Comcast bandwidth usage meter which we have been hearing about for over half a year now, I will believe it when I see it. And I am certainly not seeing it now.

Additional IPs

[XanC]

There will be no paying extra for additional IPs. Everybody will get a /64. Look at this:

Addresses available in IPv4: 4,294,967,296

Addresses available PER CUSTOMER for IPv6: 18,446,744,073,709,551,616

This enables stateless autoconfiguration (usually based on MAC addresses) that simplifies everybody's lives.

Re:Additional IPs

[Anonymous Coward]

18,446,744,073,709,551,616 IP's ought to be enough for anybody

Re:Additional IPs

[amorsen]

/56, not /64. Otherwise the customer can't subnet without losing autoconfiguration. Preferably /48.

Re:Additional IPs

[paul248]

If your ISP only gives you a /64, you'll only be able to run one subnet. IANA suggests giving everyone a /48:

http://www.iana.org/reports/2002/ipv6-allocation-policy-26jun02 (section 2.7)

But I think some ISPs are planning to do something in between.

Re:Additional IPs

[thegameiam]

Unfortunately, if you want DNS server addresses, you need DHCPv6 as well as SLAAC.

Re:Additional IPs

[j h woodyatt]

Well, I don't think you should be terribly surprised if Comcast decides you should be willing to pay extra (and agree to different terms of service) if you don't want your IPv4 address(es) to be walled off behind their Carrier Grade NAT.

Re:Additional IPs

[vanyel]

Everybody will get a /64

Actually, the policy from ARIN recommends everyone get a /48, i.e. 64K *networks* (with, for practical purposes, unlimited addresses per network, one change of mindset is to stop thinking in terms of addresses, but in terms of networks). At the small regional ISP I work for, we've got IPv6 up in experimental mode now and starting to plan for full support --- with this kind of space and some other factors of ipv6, it makes some aspects of life a lot nicer (particularly the breathing room), others a little more complicated (namely making sure your firewalls are all setup).

As for "we've been running out for 10 years", if it weren't for NAT, we *would* have hit the wall already, as it is, it's about 2-3 years out (though more intensive use of NAT could push it out more, but that'll be a royal pain for those doing it) and much of the traffic on the ARIN lists for some time now has been preparing for that end.

Another factor is that people *are* starting to deploy, and if you don't support IPv6, you're slowly going to find places you can't get to, though it'll likely be a number of years before that's significant.

Re:Additional IPs

[Anonymous Coward]

Everybody loves stateless autoconfig until they want to hit their favorite web site and realize they don't know its IP address.

DHCPv6 is just as necessary for IPv6 networks as DHCP is for IPv4 network. No big improvement there. Now you can get your IP via stateless autoconfig and your DNS servers via DHCP; in the IPv4 world, we only had to use one protocol, DHCP, to get both.

Also, I don't need 2^64 publicly-accessible IP addresses. Most people don't even want 1 publicly-accessible IP address. For 99% of subscribers, there is no legitimate Internet-initiated traffic they want to receive.

IPv6 was a neat idea 15 years ago. It is looking like a silly idea in search of a problem now.

Re:Additional IPs

[rdnetto]

There will be no paying extra for additional IPs.

You mean there will be no reason to. Chances are, Comcast will increase the cost of IPs under the excuse of recouping the costs of the switch.

[citation needed]

[XanC]

[citation needed]

Really? I wonder...

[kenp2002]

This would have NOTHING to do with monitoring and shaping your network traffic. None at all. ISPs don't do that.

And they won't be sending you:

"We have observed an unusual amount of encrypted traffic originating from your IP address" email implying that using encryption will get you disconnected.

Nope never will happen. They won't be injecting packets either to kill you VPN connections because that can't figure out what traffic you are sending. They would never do that, at least until your employers get involved asking why they were tampering with a secure connection to a financial institution. Nope not at all. Hamachi works great when it doesn't mysteriously die...

And they'll never send you a "Friendly Reminder" warning that using Tor to hide software piracy is still illegal, even if you are chatting with people in China on the annaversary of Tieniman.

Because they never inspect your traffic in order to identify what you are doing on their connecition.

They also don't send "friendly reminders" when you use PGP encrypted email that they are simply checking in on "unusual activity on their email server."

Nope, no motivation at all for switching on and using IP6 except perhaps the ability to assign static IP address for better tracking...

I wonder: Anyone out there with a brand new shiney IP6 address try a release\renew to see if you get a new address?

Re:Really? I wonder...

[AndrewNeo]

They can track your IPv4 DHCP allocations just fine. Their DHCP servers are the one that assign the IP to your cable modem's MAC.. the hardcoded, whitelisted MAC address they have in their system to give you service. They know exactly who you are right now.

Re:Really? I wonder...

[amorsen]

I wonder: Anyone out there with a brand new shiney IP6 address try a release\renew to see if you get a new address?

Your comments imply you don't know very much about IPv6. Practically noone plans to do stateful DHCP for IPv6. Release/renew won't do a thing, because the address is generated by your computer, not the DHCP server. You can change the address by changing the MAC address of your NIC or simply by picking any address you want in that subnet.

Tracking is dead easy with IPv4. Modern high-performance deep packet inspection can do practically anything you imagine except decrypt encrypted traffic (but it can do traffic type analysis even on encrypted traffic). Making the DHCP server tell the analysis servers (or the inspection engines) which IP addresses match with which MAC addresses at which times is trivial.

Re:Really? I wonder...

[oasisbob]

Your comments imply you don't know very much about IPv6. Practically noone plans to do stateful DHCP for IPv6.

Comcast, for one, does plan to use stateful DHCPv6. See this presentation from NANOG 46 for more. (Ironically, the very story we're discussing comes from Comcast's announcement at NANOG, including plenty of technical details. ) One must remember that DHCPv6 can also be used for prefix delegation, something an ISP has a need to do for each subscriber.

Release/renew won't do a thing, because the address is generated by your computer, not the DHCP server.

Perhaps, but once a user has a IPv6-capable CPE, the IP address could likely change upon RELEASE/RENEW. (See RFC 4941) You're right that once a subscriber receives a prefix, they can do whatever they want with it, and perhaps many will end up using SLAAC. But it's really difficult to say without a bunch of consumer devices on the market, and DHCPv6 is needed in most environments to augment SLAAC anyways. Unless you have a crystal ball as to how CPE devices will behave, I think you could perhaps be a bit more educational in your responses instead of berating someone who is possibly misguided, but genuinely curious.

tracert in no time at all - beat that sucker !

[Anonymous Coward]

Tracing route to ipv6.l.google.com [2001:4860:b002::68]
over a maximum of 30 hops:

    1 Destination host unreachable.

Trace complete. Gotta love that AT&T speed.

This is what it'll take for IPv6 to happen.

[IGnatius T Foobar]

I'm happy to see this. If the major ISP's start rolling out IPv6 to customers, then we'll really start to see the chicken-and-egg deployment problem get solved. In the US there are really only half a dozen of The [Phone|Cable] Companies that need to get on board to cover the vast majority of Internet users.

Stupid

[scubamage]

First, IPv6 is still a draft standard(s) to my knowledge. Many pieces of equipment aren't interoperable because of conflicting draft standard revisions. Further, the IPv6 stack gets updated in windows updates, and suddenly everything is broken. We have had this happen for a bank who tried upgrading to IPv6. The deployment went smoothly, until a windows update changed the IPv6 stack to use a different standard from the standard being used by the networking hardware. Suddenly they lost connectivity with all branch offices and had to pull back the update. A day's worth of productivity ruined because of this. Further, how are they going to solve the other issues with IPv6? Dual stack, teredo tunneling - none of these things are standards. They could handle all of it on the network shy of the last mile with teredo tunneling, but then the clients are still limited to IPv4 addresses. This is like someone saying "hey, we want all of you to use this thing we're not sure will work yet." It's foolhardy. Let the IEEE do their work and roll crap out when it's finished. Using your paying customers as beta testers is foolish - nay - freaking retarded.

Re:Stupid

[kimvette]

Using your paying customers as beta testers is foolish - nay - freaking retarded.

Hey, it's worked for Microsoft for over 30 years now!

/obligatory

Re:Stupid

[Cyberax]

IPv6 core is NOT a draft standard for about 8 years now.

Some details were only recently approved, but they generally do not have effect on end-user (Windows XP already had a working IPv6 stack).

I've been running IPv6 on my computers (including Windows XP-based ones) for a year now, without any problems.

Re:Stupid

[Dahan]

IPv6 was published as a standard in December 1998; it's definitely long past the draft stage. And the IEEE doesn't have anything to do with IP standards.

Re:Stupid

[Azh Nazg]

Y'know, IPv6 was a finished standard 10 years ago. IPv6 is RFC 2460, dual stack is RFC 4213... The *only* thing you've named that's even a draft standard is Teredo. Otherwise, IPv6 was recommended for general use and roll-out years ago. That Windows Update is breaking things is a case of your IT staff being incompetent, not IPv6 being a "draft standard" or "not ready for use".

Re:Time Warner is already doing this in Brooklyn/N

[quazee]

That's because you are using an IPv6 address in the 6to4 address space, not a native IPv6 address.
And according to trace, your ISP doesn't have their own 6to4 router deployed, so the traffic gets sent to whoever announces the shortest route to 192.88.99.1 route via BGP.
(192.88.99.1 is a special IP which means 'any 6to4 router')

Re:Time Warner is already doing this in Brooklyn/N

[Movi]

Nope, not really

1  [My IPv6]  1.421 ms  1.087 ms  2.245 ms
2  [My Tunnel]  35.730 ms  38.181 ms  34.940 ms
3  gige-g2-4.core1.fra1.he.net  33.940 ms  34.452 ms  33.944 ms
4  de-cix20.net.google.com  45.923 ms  43.556 ms  39.865 ms
5  * * *
6  fx-in-x68.google.com  56.283 ms  50.369 ms  36.717 ms

But there's so much space for stuff like...

[istartedi]

...IP addresses that spell things out with the available characters and number.

When I was messing around with the tunnel brokers a few years ago to develop some stuff that was supposed to be IPv6 ready, I saw plenty of addresses that had dead:feed and of course, the ever popular dead:beef in the logs.

Besides, how often do you put IPs in anyway?

If you absolutely must use an IP, of course you still need to remember the subnet, but after that it's a blank slate for your mnemonic license-plate style amusement.

Re:But there's so much space for stuff like...

[Anonymous Coward]

dead:beef?

Re:But there's so much space for stuff like...

[dennypayne]

And "for all intensive purposes" was never the phrase. It's "for all intents and purposes."

Where is Mark Lottor? IPV4 has plenty left to it!

[aisnota]

The large telecoms and cable outfits have tons of unused IP space that could be CIDR blocked out, think of the class A 24.X.X.X for instance that used to be @Home and Rodgers, large portions are empty! AT&T moved @Home to 12.X.X.X and then subsequently provides managed space to cable outfits like Mediacomm etc.

Now Mediacomm has just finally got around to getting its own space, is AT&T offering to CIDR out their precious class A?

No of course not, like some of the others, they get allocations from ARIN and sit on them instead of consolidating. They have scads of CIDR blocks used by all sorts of companies out there. Heck ARIN should just re-map some of those AT&T direct to the customers, let them keep the 12.X.X.X A Space.

Back in the day, Mark Lottor did mapping of all live ping able IP's before firewalls were so common and NAT extremely rare. If he were to make a comparison with whomever does like mapping today to those legacy maps and IP allocations, it would be a fascinating graphic to show the transformations and if by carrier, show how greedily the Worldcom/UUNets Sprints and Baby Bells have asked for space, color to their identity and now look to see many time those scattered CIDR blocks are empty. Sprint, old UUNet and Baby Bell CIDR's if unused, should get back into the pool.

Where is Mark Lottor and these newer guys with the latest IPV4 utilization's mapped out for the comparison analysis.

Enough said.

 

Lots of ISPs already have IPv6, including Verizon

[XMLsucks]

Lots of American ISPs are already providing IPv6 because they want to have the government as a customer. Many of you probably could enable IPv6 but don't because your router is incapable of handling IPv6. There are very few home routers that I could find that support IPv6. One that does is Apple's Airport Extreme. I bought that, connected it, and instantly got IPv6 addresses handed out to my home network. Although they are 6to4 addresses, I can connect to other IPv6 hosts, including friends at other ISPs, and ipv6.google.com. When I'm remote, I can connect directly to any of my home computers (when using IPv6) --- no more port forwarding via NAT. One reason that 6to4 appeals to the ISPs is that it puts a time limit on your IPv6 prefix lease which is tied to the lease on the IPv4 address. Thus when the IPv4 address changes, your IPv6 subnet's prefix changes, which makes it hard to run a server, and you must rely on dyndns. Dyndns with IPv6 is very easy, because your end host knows its IPv6 prefix (and doesn't have to ping a remote host to figure out its IP address as is necessary for a IPv4 host behind NAT), and because everything on your subnet knows instantly when the IPv6 prefix changes, and so you can update the dyndns with a very small race condition.

That is great.....

[wpiman]

but when are they going to bring a stable version of their IPV4 offering. My connection goes down quite often. And their TV boxes are even worse.

IPv6 = no NAT? Not if Comcast has its way.

[glindsey]

When Comcast switches to IPv6, do you really think they'll give you more than one IP address? You better believe they'll charge you more for each additional one. Maybe they'll give you two or three for free, but I doubt it.

So unless you want to pay per computer you have connected, you'll still need to NAT them through a router. Nothing will change.

Re:Time Warner is already doing this in Brooklyn/N

[thesandbender]

That's because you're using your own tunnel... not there's. If I setup a hurricane tunnel on my router than I would have the same trace. My comment centers around the fact that this is Time Warners *default* behavior. So, as more users start to use IPv6 aware apps there will be increased traffic going to the gateway that TW is using in Amsterdam... which is silly.

Re:IPv6 = no NAT? Not if Comcast has its way.

[DaMattster]

Considering that I've always considered Comcast, the "Microsoft" of the ISP world, I likely to agree with you on that one. More IP addresses means more money to charge and more money to pocket. I am not skeptical, I am simply looking at the nickel and dime scheming going on. Much like the recent shake up in the wireless telecom industry, we need one in the ISP one as well. For those of you who don't know, the shake up in wireless telecom was Boost Mobile's 50 dollar unlimited plan which forced many carriers to lower prices to compete. We need a Boost equivalent of an ISP.

$5 - $6 PER IP / SYSTEM LIKE HOW BILL YOU PER CABL

[Anonymous Coward]

$5 - $6 PER IP / SYSTEM LIKE HOW BILL YOU PER CABLE BOX / CABLE CARD / OUTLET FEES?

Must use the comcast router that has NAT locked out so you are forced to buy a ip per system? or have to only buy 1 ip use 2 routers.

They are ripping people off on the cable side with outlet fees and they now want to move that to there cable internet side what is next pay $3 per phone on there cable phone line like how ATT used to bill you?

OMG, someone made me laugh..

[geoncic]

Very rarely can I say "LOL" and mean it! That is just gold... thank you

huh?

[XanC]

RFC 2460 was publish in 1998. There's nothing "draft" about IPv6; it's quite mature. Sounds like you have a Microsoft problem, not an IPv6 problem.

No DHCP

[mdmkolbe]

IIUC, with IPv6 you don't have to run a DHCP server that keeps track of all assigned addresses. Instead you just have a server that periodically announces the network's link address.

But I think you are right. From the user's perspective, IPv6 won't change anything(*) just like from the user's perspective moving from 16-bit to 32-bit didn't really change anything.

(*) The one killer-app I've heard of for IPv6 is IPSec, but SSL, TLS, IPSec-on-IPv4 have kind of taken the winds out of that sail.

Anonymous Coward

[Anonymous Coward]

Compost will probably screw this up. I worked for an outsourcer for a while supporting them. Never in my life have a seen such a disorganized mess as Compost. Due to the fact that different markets were completely on their own with absolutely NO standarization, any time they attempted any sort of change the result was something that ended up not working and customers fighting to be compensated for downtime.

Re:Anonymous Coward

[JSBiff]

Err, I might be wrong. . . but while it's possible (may even be the default - if that's true, that's unfortunate) for your IPv6 address to use the Mac address as the last 48(?) or whatever bits of the IP address, I don't believe you *must* do that. I believe you can just use ::1, ::2, ::3, ::4, etc as the 'host' portion of the IPv6 address, can't you?

It's my understanding that IPv6 really doesn't care what the last 48 or 64 bits (I don't remember the exact number of bits for the host portion - just that it's a very large number of em) of the address is, so long as it's unique? I think the use of Mac addresses was just an 'easy' way to get a unique bitmask for that part of the IP address, isn't it?

Re:Anonymous Coward

[Anonymous Coward]

Correct. I remember seeing something about using random bits there as another option.

Re:Anonymous Coward

[Just Some Guy]

I do that in practice all the time. Our FreeBSD servers host many jail environments on each physical machine, so I use something like dd if=/dev/random bs=1k count=1 | md5 to generate the last 64 bits of each jail's inet6 alias.

Re:Anonymous Coward

[LakeSolon]

I still don't know why some people feel that the short version of Apple Macintosh needs to be typed all uppercase (MAC).

But now you've got me really confused by only capitalizing the acronym for Media Access Control.

Re:Where is Mark Lottor? IPV4 has plenty left to i

[Just Some Guy]

The large telecoms and cable outfits have tons of unused IP space that could be CIDR blocked out

No, they don't. The last I heard, reclaiming all /8 netblocks would return something like 8% of available space back to the pool. When usage is growing exponentially (or would be if it wasn't constrained to a tiny fishbowl), 8% isn't worth the aggravation.

What is the hardware impact on consumers?

[MrCool80s]

So I searched newegg.com and cdw.com for "ipv6" and with a few exceptions, only high end networking equipment results running ~$1k+ came up. Searching the web for "ipv6 hardware requirements" does not lead to much (I confess I do not want to read the whole spec.), but the article on wikipedia leads me to believe that home routers (and maybe managed switches) could be upgraded if resources such as RAM (and EEPROMs?) are sufficient and manufacturers so inclined.

Why do there seem to be so few end-user products which are or claim to be "ipv6 ready"?

Will the implementation of ipv6 end up providing every ISP account holder with a static IP, with IPv4/NAT behind it until all the old equipment dies over the next couple decades?

Re:What is the hardware impact on consumers?

[JSBiff]

Chicken-and-Egg: There is very little home networking equipment available in the US that currently supports IPv6 because there are no ISPs in the US who have deployed IPv6, so there is no market for such equipment.

However, that said, I learned awhile back (I think it may have been from another /. poster) that Apple's Airport line of home routers (at least, the current generation) support IPv6. The Airport express (WiFi-only) is $99. The Airport-extreme, which adds a few local ethernet ports on the back of the device (like a Linksys or Netgear router) in addition to wifi, is $179. Seems a little expensive, but it's cheaper than the $1k+ you were talking about earlier.

Another option is to get a router which is supported by one of the Linux-based 'alternative firmwares', and 'upgrade' one of the non-IPv6 compatible home routers to add support for IPv6. You'll probably void your warranty, but you'll get an inexpensive router (as I'm posting this, Newegg is advertising Linksys WRT54GL routers for $50), which has IPv6 support.

Re:What is the hardware impact on consumers?

[MrCool80s]

That helps my understanding, thanks.

I recall reading that the "tomato" firmware for wrt54 routers enabled ipv6 routing, but what i don't understand is how the new firmware gets around the embedded ipv4 mac...does it just spoof an ipv6 translation of the mac and be done with it that way? If the bulk of consumer devices (sold in the future) can be upgraded by rewriting part of the firmware, I thought more manufacturers than just Apple would be on their way to ipv6 for consumers. Then again, when I read router/switch specs, the "MAC address table" or similar is always listed as its own value/entity which leads me to believe it's not a trivial upgrade.

Re:What is the hardware impact on consumers?

[JSBiff]

Mac addresses aren't part of IPv4 or IPv6 - they are a different layer of networking - the physical/ethernet layer. Now, it's true that for autoconfiguration, because the Mac address is a globally unique (unless you've changed it) 48-bit value, IPv6 implementations will often use the Mac address as the 'host' portion of your IPv6 address, but far as I know, IPv6 doesn't really care what your mac address is. In both IPv4 and IPv6, a IP address gets mapped by the local router (I think, I'm not a network engineer, but this is my understanding) to a Mac address for final delivery, but that mapping is only maintained on your local network.

It's like this, I think: Someone sends a packet to an IP address across the Internet. The sending party has no idea what your Mac address is. In the transfer across the Internet, that IP packet might cross a dozen different type of physical links, each with their own addressing schemes (ATM, Frame Relay, PPP, Ethernet, etc). All that's really important at that point is the IP address. When that IP packet finally reaches your home router, your home router just has to know that IP address 123A::1 resolves to the ethernet network device with Mac address 01:23:45:67:89:ab. I think that's just maintained in a table in RAM in your home broadband routers (higher-end routers like Cisco gear might do something that is faster than a table in RAM, I dunno, but for home-use gear, where you have a very small number of devices on the network, storing those mappings in RAM should be fast enough I think).

Re:What is the hardware impact on consumers?

[Guyver3]

D-link DIR-615 hardware rev C1

no ipv6 firewall rules, but does native/static tunnel/6to4, and announces a /64 to your wired lan/wireless

IPv6 on Tomato firmware?

[Anonymous Coward]

I have a WRT54G running Tomato and Comcast gives it a IPv4, and Tomato assigns IPv6 to my internal network.

How did you get IPv6 working on Tomato? I was under the impression that it wasn't supported.

Does anyone have instructions?

tinfoil hat ++

[thegameiam]

Just because you can't ping something doesn't mean it isn't in use. ARIN and the other RIRs require extensive documentation before they give out more space, and all of the companies you've mentioned have received it. I recommend reading up on how a SWIP works, followed by getting an understanding of rWhois. At that point you might have a better understanding of some of the issues. Heck, NANOG has had some excellent discussions on the subject of IPv4 address reclamation, and the outcome of those discussions is that it's a lot of work for very, very little benefit.

s/octal/decimal/g

[Kadin2048]

IPv4 is decimal, obviously, not octal; you just stop incrementing each group at 255.

I wonder how this will integrate with local DHCP?

[wowbagger]

Like many /. readers, I have a firewall and local DHCP server, handing out addresses in one of the reserved ranges (e.g. 10.x.x.x) for my home network.

OK, so when my ISP starts handing me real IPv6 route-ability and a real IPv6 address range, how do I configure my DHCP server to take that address range and convey it to my local clients?

Yes, I know that the bottom 48 bits of the address can be the MAC address of the device, but I still need to communicate to the devices what the prefix is.

And I will *still* need that firewall and router because I have many devices that are IPv4 only, and I won't be replacing them anytime soon, so even if IPv4 vanishes from the Internet at large, I will need my firewall to proxy for those older devices.

Re:Time Warner is already doing this in Brooklyn/N

[thesandbender]

I didn't make my point clear. *All* IPv6 traffic goes through SWIP... no matter what. Even ipv6.he.net goes through Amsterdam, and HE has three IPv6 peering sites in NYC. It seems to me that if BGP was setup properly it would use an HE router instead. So I'm going to make a WAG and say either:

1. They only peer with SWIP for IPv6
2. They didn't implement BGP properly

Either way, routing all NYC IPv6 traffic through a 6to4 router in Europe doesn't make sense.

Anonymous Coward

[Anonymous Coward]

I would beg to differ. Do to all the nat routers being used in homes and business the so "called running out if IP's " is a not issue in the near term 10 or so years. I would also hope people realise that IP v6 has a large privacy issue. in like now the IP does not hold the mac address in the packet .. IP V6 will. So not only will "three letter government" divisions have the ip address of your machine even if you change they will have the mac address as well.

Think about that be for rushing headlong into ipv6,,

Re:Time Warner is already doing this in Brooklyn/N

[BitZtream]

Lets clear this up.

All you know from this traceroute is that the routers between 2002:185a:90f:1234::1 and ams-core-1.tengige0-0-0-0.swip.net are acting in a transparent manner. It could be because they are not decrementing the TTL on each hop.

This could be because they are transparent routers, it could be an IPv6 tunnel over IPv4 or something else, you really don't know and are making silly assumptions.

What bothers me however is that either your Windows Vista/7 PC (as noted by the C:\Users in the command prompt and your use of windows tracert instead of traceroute) is directly connected to the Internet, while it is possible that you are doing that, it would be utterly stupid and I'm going to make an assumption of my own, that you are not directly connected to the Internet. Why do I make this assumption? Well partially because its a rather quick way to get exploited, theres always SOMETHING you can exploit in an MS OS and that it means you only have one PC, being that this is slashdot I can guess that those are not the case, so you aren't directly connected to the Internet and the first hop you're talking to is a DLink or Linksys router or something.

Now this makes sense, as it simply means your router is connected to swip.net using an IPv6 over IPv4 tunnel. Since this is a free service and several consumer grade devices support it, this is more likely the case. I'm not real sure how you end up with IPv6 enabled on your router and not have any clue about it, but perhaps it was done by a roommate or something like that.

Eitherway, me thinks it might be better for you to learn wtf is going on with your own internet connection than talk about how Time Warner handlers theirs.

Finally, since you're obviously new to IPv6 and networking. SWIP sells connections, they are a backbone provider which is why you see a direct connect from them to Google. They also provide IPv6 tunnel endpoints so you can tunnel it over IPv4, which appears to be exactly whats going on in your case. This tunnels are free to anyone who signed up. With that in mind and the fact that tunnels have to generally be setup on both ends in advance its likely that if Time Warner IS involved in this, they are simply working a deal with SWIP, not robbing service from them. I would have to say that SWIP.net is fully aware of the tunnel route and has authorized it, that is after all one of their core businesses.

I suggested you learn a little more about the current state of IPv6, the existing providers with IPv6 support, and most importantly, what your little Linksys or DLink router is doing that you are completely unaware of. At least go turn off your tunnel to swip.net before claiming that TWC supports IPv6 in your area.

Corrected explanation...

[Kadin2048]

Just to clarify the bit about the addresses, because I forgot a couple of sections...

The whole IPv6 address is 128 bits, but in a unicast address, the first 3 are the "prefix identifier," basically saying that this is unicast. Then you have a 13 bit "TLD Identifier" and 8 reserved bits, completing the global prefix portion.

But then you have a 24 bit "NLA ID", which might specify an ISP or some other intermediate network. This provides for traffic aggregation, and they get assigned (I guess) by the national registries. This brings you to 48 bits. Exactly how they'll choose to distribute the NLA IDs, and how many each organization/ISP will get, I'm not quite clear on. I've heard some people allude to ISPs getting large blocks at this level and putting a "subscriber ID" or "customer ID" in this region, leaving 80 bits free per customer, but I don't think this is really the case.

After the NLA is a "SLA ID", which is like a very big subnet identifier. It's 16 bits long, bringing you to 64 for the address so far. This is what I think individual home routers will get from ISPs, assuming the NLA IDs get given out with enough granularity so that there isn't competition.

Beyond the SLA ID is 64 bits for the "interface ID," which a host can pretty much define however it wants. In most applications this can be easily created by padding out the Ethernet MAC, although it can also be generated randomly if that's not desired.

References:
http://technet.microsoft.com/en-us/library/cc757359(WS.10).aspx - Surprisingly good TechNet article
RFC 2462

Re:Where is Mark Lottor? IPV4 has plenty left to i

[BitZtream]

Uhm, reclaiming ALL /8 netblocks would return 100% of the network.

You people really need to get over this classful idea of routing and assignment, its hasn't been that way for years, we use subnet masks now, you heard of them?

But back to the errors in your statement, each /8 assignment accounts for approximately 3.125% of the total network address space (not usable space, TOTAL space). So about 2 and a half of these assignments now account for your 'random, pulled out of your ass 8%'.

Well, since I know that it is impossible to get anything close to 8% of the address space using only /8s I can immediately assume you don't have a clue at all.

What else? Well, there are 128 /8 blocks in classful routing. That is half the address space total, or 50% if you recalled them all and ignored other classful networks. Again, we're not seeing your 8%.

So, if it isn't possible to get 8% of the address space using /8 allocations, perhaps you should consider that the source of said information is most likely incorrect. If they can't do simple math, its going to be hard to take anything else seriously. When you start a statement, as if it were fact, by saying 'last I heard' its generally a sign that you are wrong and someone is going to point it out to you shortly.

For more clues about classful, classless and the Internet back when EVERY assignment was a /8 by design, please checkout the following Wikipedia articles, they make great starting points to get a clue.

http://en.wikipedia.org/wiki/Classful_network
http://en.wikipedia.org/wiki/Classless_Inter-Domain_Routing
http://en.wikipedia.org/wiki/IP_address_shortage

Is reclaiming the /8s worth it? Yes, far more so than reclaiming the smaller wasted blocks. But thats the problem, blocks were allocated poorly from day one and now its a freaking mess thats a bitch to clean up. So we have two choices, clean it up or move to IPv6.

In case you haven't noticed, the current trend is to just clean it up, much to the disappointment of many slashdotters who want IPv6 so bad it made them forget why exactly they wanted it in the first place. :)

Re:Where is Mark Lottor? IPV4 has plenty left to i

[Just Some Guy]

Uhm, reclaiming ALL /8 netblocks would return 100% of the network. You people really need to get over this classful idea of routing and assignment, its hasn't been that way for years, we use subnet masks now, you heard of them?

That's what we call "irony". You see, there aren't that many /8 netblocks, and you don't hear people clamoring for the subnetting of anything smaller. I mean, what are you going to do with a /20? Break it up and route two /21s?

But back to the errors in your statement, each /8 assignment accounts for approximately 3.125% of the total network address space (not usable space, TOTAL space).

There are 256 /8 netblocks, each accounting for about .4% of the TOTAL space. If you somehow missed that, then you're not really qualified to argue either side of the debate.

Re:I wonder how this will integrate with local DHC

[JSBiff]

To me, that's the most interesting aspect in watching to see how this goes for Comcast - I imagine that, at least at first, they will run *both* IPv4 and IPv6 on their network. I imagine for existing customers who aren't interested in upgrading, it'll still be IPv4 for quite a while.

But, Comcast has an opportunity, with new customers, to start deploying IPv6. The trick here is, from the customers' perspective, their local network could still support IPv4, I suspect. If you check my journal, I had posted an article I wrote up some months back, describing how I think someone could create a router device which allows IPv4 devices on the local network to use IPv4 locally, but have an IPv6 address which the outside world sees. Basically, the router would do mapping and translation between the (global) IPv6 and (local) IPv4 address.

All of the necessary 'logic' to get this to work could most likely be bundled into a firmware on a device like a Linksys WRT54GL router. Any computers or devices which understand IPv6 could use IPv6 locally, while any devices which don't understand IPv6 could continue to use IPv4 on the local network.

It'll be interesting to see if Comcast gives customers who trial the IPv6 connections, such a router device to take care of all this stuff.

The problem with IPv6 is...

[Anonymous Coward]

Where are we going to get a database to handle the DNS requirements? Who's going to reward all the names people will give their toasters?

What about privacy and tracability?

[Anonymous Coward]

O.k., I am not an expert in these things, but it seems to me that IPv4 needs a lot of address translation (routers instead of simple bridges). Doesn't this mean that, when viewed from the Internet, your PC is more secure, because only the external node, i.e. the cable modem, is visible from the outside? This way, a connection can mostly be established from the inside, but an outsider trying to snoop your LAN will face several hurdles.

Now compare that to IPv6: That's like having a static IP for every single device in the world. Every packet is *totally* traceable. Why is it that privacy-conscious people are not rallying against IPv6????

Please elaborate.

Re:What about privacy and tracability?

[JSBiff]

1) Privacy: IPv4 addresses can already, largely, be traced back to at least your ISP account, if not the specific device on your local network. You have no IP address privacy. Go read about the Jammie Thomas / RIAA case if you doubt it.

2) Security: NAT provides, basically, a type of a firewall. Firewalls existed before NAT. There's no reason your 'home router' which currently does NAT couldn't firewall your connections in a very similar way to how NAT protects you currrently (e.g., only outbound connections are allowed by default, and in-bound connections when you purposefully open a port). NAT does not give you any better security than a firewall.

What it does do is make it a pain when you *want* to allow people to connect directly to your computer for things like games, Voice- or Video-chat, direct file transfers, bittorrent, etc. This is particularly if you have more than one computer you want to allow people to connect to (yes, with NAT you can do a port forward, but that only works for a single computer; if you have 2 or 3 computers, your pretty much screwed).

But, you say, the Internet works fine right now and I have NAT, but I have no problem using Skype, games, bittorrent, etc. That is because, in order to deal with everyone using NAT, pretty much everyone has come up with proxy schemes. What do I mean by that? Instead of two computers connecting directly to each other, when a NAT is 'in the way', typically the two computers will connect to a third computer. This works, BUT, it's inefficient - it means that your packets have to route through a longer chain of Internet connections (usually), and if the proxy server does not have sufficient bandwidth, your throughput is limited by that proxy server.

So, in the bittorrent example, I believe that if you join a torrent, and you are behind a NAT, then in order for you (A) to upload a chunk of data to another user (B), you both have to connect to a third user (C) who is NOT behind a NAT (or at least, who has properly setup port-forwarding, and I suspect a great many bittorrent users don't really know how to setup port forwarding correctly). If that NAT wasn't there, the torrent would be faster for you and everyone else who is part of that torrent.

The Internet works with NAT, but it would work BETTER without NAT.

Re:What about privacy and tracability?

[Anonymous Coward]

I am still not convinced. Say I am in Iran and using some unsecured wifi network (let's assume at the University) to post a regime-critical anonymously to slashdot. I can feel somewhat safe, can't I?

But now imagine I have an IPv6 assigned to my laptop - essentially a hard-coded serial number. Because let's face it, isn't this where all this is going eventually?

And yes, at that point no matter where I go, I can always be identified.

Re:What about privacy and tracability?

[ion.simon.c]

a) IPV6 autoconfig allows you to optionally choose a random address within the subnet that you've been handed. Your address doesn't have to be hard-coded and can change periodically.
b) You're looking for security and anonymity in the wrong place. IP wasn't designed to provide anonymity. You need to focus on protecting the contents of your transmissions and/or obfuscating the details of who you are talking to. Look towards using something like IPSec or Tor.

Do you even know how DNS works?

[JSBiff]

DNS isn't a single centralized database. It's many, many databases, organized hierarchically. Granted, if the Internet keeps growing, I suppose there could still be some scalability issues.

See this article for an explanation of how the DNS 'database' is broken up into pieces which are handled by different servers.

Re:I wonder how this will integrate with local DHC

[Cajal]

They will possibly use DHCPv6 Prefix Delegation to configure your home router's DHCPv6 server. DHCPv6-PD is one of the more useful aspects of DHCPv6.

Also, Comcast will deploy IPv6 alongside IPv4, so your IPv4-only devices will continue to work just fine.

Re:Time Warner is already doing this in Brooklyn/N

[quazee]

Anything from an 6to4 address typically gets routed to 192.88.99.1 (IPv4, protocol number 41), unless IPv6 is configured in a really weird way.
Since your ISP does not have their own router with the 192.88.99.1 anycast address, *all* IPv6 traffic goes through one of their peers who advertises their route to 192.88.99.1.
The actual destination IPv6 address doesn't matter (unless the destination is also a 6to4 address, in which case, the traffic is typically routed directly to the encoded IPv4 address instead of 192.88.99.1).

Re:IPv6 = no NAT? Not if Comcast has its way.

[JSBiff]

Funny thing about Boost is that they are owned by Sprint. I'm left scratching my head wondering why Sprint offers a $99 dollar unlimited plan while another part of the same company is offering virtually the same plan for $50? I guess there might be one important difference - the Boost plan offers unlimited 'web', but not ulimited 'data' which probably means you can't use something like a Blackberry or other smartphone with apps. So, no AIM or Skype unless you want to cough up another $49/mo.

I also have to wonder how long before the corporate overlords at Sprint kill this deal? Might just be a somewhat short term offer to grow the brand.