One Step Closer to IPv6

gbjbaanb writes "IPv6 came a step closer yesterday as ICANN added IPv6 host records to the root DNS servers, reports the BBC. 'Paul Twomey, president of Icann which oversees the addressing system, told the BBC News website there was a need to start moving to IPv6. "There's pressure for people to make the conversion to IPv6," he said. "We're pushing this as a major issue." The reason for the urgency, he said, was because the unallocated addresses from the total of 4,294,967,296 possible with IPv4 was rapidly running out. "We're down to 14% of the unallocated addresses out of the whole pool for version 4," he said. Projections suggest that this unallocated pool will run out by 2011 at the latest.'"

Comments

Sad

[suso]

Its sad to look at the list of class a allocations and know that we're almost out. All this was done before NATs became popular. I think ICANN/IANA should work on wrestling some of those class As back from companies like Ford, Apple, HP, etc. None of those companies are going to ever have 16,000,000 hosts on public IPs. I know some of those companies have already made sub allocations. We could probably buy 5-10 years if they could reclaim just the 3, 9, 13, 17, 19, 20, 34 and 40 class As and get over 130,000,000 IPs back.

I mean, if those companies complain, who cares. They wouldn't get such large and prestigious allocations in an IPv6 network anyways. So what's the difference.

I know, I know, we should move to IPv6 anyways. Just a suggestion. Poor initial planning warrants changes down the road.

Re:Sad

[Anonymous Coward]

I completely agree with you. That is why I am going to pledge my entire allocation of the 10.0.0.0/8 network back to the IANA. As long as we ensure that it is reallocated properly, I think it will be a huge benefit for the Internet as a whole. For those of you who might control a part of or the entire 172.16.0.0/12 or 192.168.0.0/16 network, you might want to ask yourself this question: do I really need that many addresses?

Re:Sad

[Profane MuthaFucka]

130x10^6 addresses isn't that many. It'll push off the exhaustion of the address space by a year or two at the most, and then we're still going to need IPv6.

Also, without IPv6, there's only a maximum of 2^32 Linksys routers that will be needed. IPv4 is unfairly capping the maximum number of needed NAT routers, and thus unfairly capping the profits of Cisco. We must think of the cost of IPv4 in terms of corporate profits, or we are doomed. Our economy depends on exponential growth, and that applies to addresses on the Internets too.

Re:Sad

[93 Escort Wagon]

Its sad to look at the list of class a allocations and know that we're almost out. All this was done before NATs became popular. I think ICANN/IANA should work on wrestling some of those class As back from companies like Ford, Apple, HP, etc. None of those companies are going to ever have 16,000,000 hosts on public IPs. I know some of those companies have already made sub allocations. We could probably buy 5-10 years if they could reclaim just the 3, 9, 13, 17, 19, 20, 34 and 40 class As and get over 130,000,000 IPs back.

I mean, if those companies complain, who cares. They wouldn't get such large and prestigious allocations in an IPv6 network anyways. So what's the difference. Halliburton has a TLD? I dunno... if you tried to take it back, a certain someone might "accidentally" shoot you in the face!

Re:Sad

[Anonymous Coward]

Your point is well taken, but don't hold your breath. An IPv4 allocation is an asset, and the bigger the allocation, the bigger the asset. Don't expect corporations to give away their assets without expecting something in return. Just sayin'

Re:Sad

[tknd]

We could probably buy 5-10 years if they could reclaim just the 3, 9, 13, 17, 19, 20, 34 and 40 class As and get over 130,000,000 IPs back.

130,000,000 / 4,294,967,296 = 3%

The article says we will run out of unallocated IPs by 2011. The unallocated pool is 14%. It is currently 2008. 2011 - 2008 = 3 years. What makes you think that reclaiming 3% is going to buy us 5 to 10 years?

Re:Sad

[93 Escort Wagon]

Er, sorry, not enough caffeine. Definitely NOT a TLD, but they do own the 34.* block.

Sheesh.

Re:Sad

[Rolgar]

All class A's should be re-designated as class B's, and entities that currently have class As that need more than a class B should be able to claim multiple class B's from their current class A.

I'm a contractor with the Postal Service (Class A 56) and I don't think we need the whole thing. Probably 50-75% of postal computers are individual post offices that access the network through a DSL (or in some small towns, dialup) and VPN. Data Centers and other large facilities should easily be able to fit in 1-10 class B's depending upon just how many sites there are.

Re:Sad

[KanjiMonster]

I mean, if those companies complain, who cares. They wouldn't get such large and prestigious allocations in an IPv6 network anyways. So what's the difference.
Not relatively, but absolutely.
Plan is for *consumers* to get a /48 allocation, thats 2^80 IPs. Thats 2^48 Internets! ISPs will get from /22 to /19, dependent on their size.

Re:Sad

[TheRaven64]

Thats 2^48 Internets! Why would you want that? One of my staff sent me an Internet the other week, and it took three days to arrive! If everyone has 2^48 Internets, my staff's Internets will never arrive.

Re:Sad

[misleb]

Didn't Bill Gates once say, "127.0.0.1 should be enough for anybody." Damn, he's always so short sited.

Re:Sad

[Arterion]

I wish I had mod points. I got strange looks in the office from laughing out at that comment. :)

Re:Sad

[gbjbaanb]

I mean, if those companies complain, who cares. They wouldn't get such large and prestigious allocations in an IPv6 network anyways. So what's the difference. yeah, we'll restrict them to a meagre 281474976710656 addresses like everyone else. That'll show them, and if they want more than a single /48 then they can just go whistle, loooosers.

Re:Sad

[Above]

The current IPv4 burn rate is around 10-11 /8's per year worldwide. You list 8 /8's that might be recovered; that would be under 1 year of additional time at the current rate.

For more information on the rate of IPv4 consumption, see http://www.potaroo.net/tools/ipv4/

And, while those companies aren't using 100% of their blocks, they are using some of them, so it's not as trivial as just returning the unused block; they have to make sure they are numbered in a small subset of it and return the rest.

A lot of effort to go through to delay the inevitable by under a year.

Re:Sad

[Guppy06]

I just hope that the guy who holds the 127.0.0.0/8 network never follows suit. All his hosts have the largest pr0n collection I've ever seen!

Re:Sad

[Cheeko]

HP actually has 2 Class A's and a class B (though they may have let the B go). When I was working there (till this past year), everything we did generally had a 15 or 16 IP. (15 = HP, 16 = DEC, the class B was the old Compaq)

Re:Sad

[CopaceticOpus]

But 3% of 2011 is over 60 years!

Re:Sad

[MikeyTheK]

We've been hearing about how we were about to run out of IP4 addresses in three years back in 1990. It's been 17 years. The same silly article gets published every six months. I'm still waiting. Call me when we actually hit Peak Oil, too. Oh wait! We just passed it AGAIN!

Re:Sad

[A beautiful mind]

I challenge you to bring 5 mainstream articles each six months apart from the past 10 years.

Re:Sad

[houghi]

I hope nobody messes with those IP adresses, as I have hackme.houghi.org pointing to one of those adresses.

It is a bitch to get a reversed though.

Re:Sad

[burndive]

I believe you mean short cited.

Re:Sad

[WhiteDragon]

We could probably buy 5-10 years if they could reclaim just the 3, 9, 13, 17, 19, 20, 34 and 40 class As and get over 130,000,000 IPs back. Well, think about this: 3.x.x.x is owned by General Electric:

whois 3.0.0.0

OrgName: General Electric Company

NetRange: 3.0.0.0 - 3.255.255.255
CIDR: 3.0.0.0/8
NetName: GE-INTERNET

So naturally, you would expect www.ge.com to be in that block. And you would be wrong.

dig www.ge.com ;; QUESTION SECTION: ;www.ge.com. IN A ;; ANSWER SECTION:
www.ge.com. 30 IN A 216.74.131.56

I have always thought it was rather irresponsible of them.

Re:Sad

[WhiteDragon]

I'm a contractor with the Postal Service (Class A 56) and I don't think we need the whole thing. Probably 50-75% of postal computers are individual post offices that access the network through a DSL (or in some small towns, dialup) and VPN. Data Centers and other large facilities should easily be able to fit in 1-10 class B's depending upon just how many sites there are. Well, it used to be that all that ip space was divided up amongst different processing & distribution centers, post offices, etc. Now like everyone else, they are using a lot of NAT.

Re:Sad

[BitZtream]

How do you define prestigious? I hope its not by number of addresses assigned.

I have a /64 assigned to my house. In IPv4 speak, that means my house has approximately 4.2 billion IPv4 address spaces to itself. I'll probably tunnel one IPv6 address space worth to my iPhone, and one to my car, maybe one to my toaster ... Not sure what I'm going to do with the rest. /MSG me for trade

Re:Sad

[Anonymous Coward]

I mean, if those companies complain, who cares. They wouldn't get such large and prestigious allocations in an IPv6 network anyways. So what's the difference.

Actually, they would get such a large allocation in an IPv6 network. So would everyone else.

Re:Sad

[cheater512]

Thats weird.
If I was their sys admin I'd demand permission to use 3.0.0.x for the server farm.
Man that would be easy to remember and would look cool in pings. :)

Re:Sad

[cheater512]

Dont you mean 20 articles? :)

Re:Sad

[llZENll]

As of November 2007, a daily updated report projected that the IANA pool of unallocated addresses would be exhausted in May 2010, with the various Regional Internet Registries using up their allocations from IANA in April 2011.[8] This report also argues that, if assigned but unused addresses were reclaimed and used to meet continuing demand, allocation of IPv4 addresses could continue until 2017.

-wikipedia

Re:Sad

[Xenophon Fenderson,]

Subnet availability is only part of the problem - there's also the size of the top level routing tables to consider.

Re:Sad

[ptbarnett]

HP actually has 2 Class A's and a class B (though they may have let the B go).

HP used to have another Class B: 130.168.x.x, which it acquired along with Convex Computer. However, they subsequently gave it to Agilent when spinning it off.

Re:Sad

[ch-chuck]

Just for the record, the situation is similar to what has been happening with the RF spectrum and the fcc for ages - they are always complaining that 'we are running out of spectrum', but they only mean that it has been allocated, usually to large companies. You go out and actually scan around you find that large parts of the spectrum are barely used at all. It's like a room with all the chairs reserved, so they tell you the room is full - when actually there's hardly ever anybody in there at any given time.

Re:Sad

[Sczi]

You still get Internets? I haven't gotten one in a few years, at least in my mail. I still get one in the back of magazines every now and then.

Re:Sad

[ACMENEWSLLC]

I'm ready to begin to add IPv6 to my network. 99% of my machines can support IPV6. There is no RFC1918 private space needed with IPv6 since there is so much space. I went to allocate space, but found out that I can't;

http://www.arin.net/registration/guidelines/ipv6_initial_alloc.html

Re:Sad

[Hucko]

Lets see if he can do the 5 first. If he gets them all, we may just concede the fight to him because of the shock value.

Re:Sad

[pentalive]

Actually He probably meant "Sighted"

Re:Sad

[burndive]

Either I don't get it, or you don't get it.

His joke was a play on the fact that web sites have IP addresses (and that 127.0.0.1 would be a 'short' one); my joke was that attributing the quote that he was parodying to Bill Gates is dubious.

What was your joke?

Re:Sad

[Anonymous Coward]

Aspergers much?

Re:Sad

[idiotnot]

Might that have been for their old GEnie service?

Re:Sad

[budgenator]

That's probably one of the reasons they are dragging their feet on IP6, IP4 is a limited and valuable asset and the switch will cause billions of dollars worth of imaginary assets to disappear from the corporate books. Try explaining to the stakeholder that you asset/liabilities just went red because of IP6 and they are going to feel lied to and the law suits will start flying and for good reason.

Re:Sad

[Anonymous Coward]

He wasn't making a joke. He was pointing out the correct spelling of the word in that context.

You really need to lighten up.

Re:Sad

[Anonymous Coward]

Except the original spelling was, in fact, correct.

Re:Sad

[anticypher]

So what is stopping you?

Have you not paid your 2008 ARIN fees?
Are you not an ISP?
You can't come up with the US$35 for a /32?
How is ARIN blocking you in any way?
Are you just trolling /. as a substitute for having a life?

I don't understand your complaint. If you already have an IPv4 allocation from ARIN, getting an IPv6 allocation requires only filling out the form, sending it in, and getting your allocation. They stick the $35 onto your ARIN fee at the next billing cycle. It's even easier than getting an IPv4 allocation now.

the AC

Re:Sad

[CarpetShark]

But 3% of 2011 is over 60 years!

Your math is a little faulty there, I fear. You forgot to take into account the uncertainty regarding the actual birthdate of christ. Or, indeed, whether he was even born.

Re:Sad

[TheRaven64]

You are making the same mistake that people make when dealing with centigrade instead of Kelvin. You see, 0 CE is not actually absolute zero. Absolute zero in terms of dates is around -13.7 billion[1], so 3% gives us another four hundred thousand years, not another sixty.

[1] Or 4004 BCE or last Thursday if you use one of the less well-accepted values.

Re:Sad

[Disoculated]

Except that IP addresses aren't domain names, they are owned by ARIN and allocated to companies, and ARIN is perfectly capable of taking them back if they're unused.

Re:Sad

[Anonymous Coward]

You are truly an idiot.

Re:Sad

[Mark Hood]

Yeah, but he never has anything new...

Re:Sad

[shentino]

Actually, there sorta is.

It's called "organization local scope"

Re:Sad

[Ant P.]

As are you.

Re:Sad

[ACMENEWSLLC]

>>Are you not an ISP?
No, I am not an ISP. I have 3 Internet connections & utilize BGP with my IPv4 space.

If I'm going to setup IPv6, I'm setting it up with proper addressing. Meaning address space that I can use internally & when IPv6 becomes more viable on the Internet, I can continue to use the same addressing without NAT. Correct me if I am wrong, but it was my understanding that IPv6 was to allow me to use routable addresses on all my devices since there is soooo much space. That I did not use NAT or private space with IPv6. I don't need IPv6, but I could set it up & begin to get it going on my network - steps towards getting to IPv6 on my Internet side too.

Here is the document I've read;
http://www.arin.net/registration/guidelines/ipv6_initial_alloc.html

>>To qualify for an initial allocation of IPv6 address space, your organization must meet the following requirements:

# be an LIR / ISP;

I am not a LIR/ISP. I am an end user.

# plan to provide IPv6 connectivity to organizations to which it will assign IPv6 address space, by advertising that connectivity through its single aggregated address allocation;

nope

# be an existing, known ISP in the ARIN region or have a plan for making at least 200 /48 assignments to other organizations within five years

nope

So am I reading this document wrong? Is it superceded, and ARIN's just failed to mention that on the document? If I am wrong and I can get IPv6 space, then I welcome the information correcting me.

Re:Sad

[Fastolfe]

I don't think multi-homing on IPv6 has been fully realized.

Consider:

1. using one IP address per uplink on your devices and publish all three addresses in DNS, or
2. use a low TTL in DNS and update DNS every time a link goes down with an alternate IP address.

If you don't like that existing sessions will fail and have to be restarted (TCP-based protocols), consider moving your services to other protocols (if possible), such as UDP or SCTP.

IPv4-style multi-homing doesn't scale well in an IPv6 world, which is why they're limiting allocations like that to large ISPs.

Re:Sad

[anticypher]

Since it took me only about 3 seconds to find this:
http://www.arin.net/registration/guidelines/ipv6_assignment.html

I can only assume you are trolling.

Where did you get your IPv4 netblock and AS number from? It's from them that you get your IPv6. If you deal directly with ARIN, then you are already an LIR.

Since you claim to be an end user, get yourself a PI block from your current LIR and start negotiating a BGP link to one of your upstreams that has v6 transit. At worst, tunnel your IPv6 BGP session to one of the tunnel providers until you can get native transit. Make sure your announcements show up in looking glasses before putting AAAA records externally on your sites.

# be an existing, known ISP in the ARIN region or have a plan for making at least 200 /48 assignments to other organizations within five years

You always answer yes to this, just about anyone should have expansion plans to have 200 new customers in 5 years. It's not like they are going to check on you in 5 years, allocations are only revoked for bankruptcies and occasionally for blocks that haven't been seen ever.

the AC

Just Like Oil

[mrxak]

Just like how when we run out of oil, solutions will come along, when we run out of IP addresses, solutions will come along. The only problem is people don't get very motivated until we're really on the edge. I don't have much hope for IPv6 for another few years yet. Still, progress is progress.

Re:Just Like Oil

[Anonymous Coward]

the solution came along, its ipv6

Re:Just Like Oil

[betterunixthanunix]

We have solutions to both problems. People just don't want to put in the time, effort, and of course, money to implement the solutions. Would you want to pay higher taxes to help subsidize the creation of bioplastics and wind power? Would you want to pay higher taxes to help subsidize an upgrade to broadband access and IPv6 use in your country?

Re:Just Like Oil

[mrxak]

How many people are using it though? Probably the same number of people using solar cars.

Re:Just Like Oil

[mrxak]

Which is why IPv6 probably won't get any real use until ISPs start charging a premium for you to get a unique IPv4 address since they don't have enough to go around.

Re:Just Like Oil

[Tony Hoyle]

If you want more than one some already do..

Mine gave me 16 for free, but it's a rare case of an actually good ISP.

Re:Just Like Oil

[Unoti]

If higher taxes would honestly go to bringing high speed fiber right to my doorstep, yes, I'd seriously consider it. I just don't have much faith in the government spending my money properly.

Re:Just Like Oil

[Anonymous Coward]

Just like how when we run out of oil, solutions will come along,

What was the solution when you ran out of oil ?

I'd like to know it. Please do not silly attack Iraq, etc.

Re:Just Like Oil

[joeytmann]

That's why ICANN is adding IPv6 to the root DNS servers. IPv6 adoption has to start somewhere, and for years everyone has been waiting for someone else to start the ball rolling. Well looks like ICANN finally got fed up and have given the ball a small push to see how far it rolls and is now waiting for someone else to give it another push to keep it rolling.

Re:Just Like Oil

[discogravy]

well, the PJ O'Rourke line ("We're waaaay out of whale oil, but instead of sitting in the dark we found other oils for our lamps") -- which is the standard American Republican line in this case -- is that we'll innovate a new way. So far the oil/energy problem has the same problem that IPv4/IPv6 have, namely that everyone wants it, but no one wants to be the one to start (or wants it enough to actually *DO* something).

Re:Just Like Oil

[Anonymous Coward]

Were you out sick the day they taught about past and future tenses?

Re:Just Like Oil

[jguthrie]

They should have left the 6bone up and had people adopt IPv6 from the outside in rather than believing that the job was finished when the backbones were sort-of IPv6 capable. Demand for addresses comes from the leaves, not the trunk.

Re:Just Like Oil

[ThreeGigs]

Well looks like ICANN finally got fed up and have given the ball a small push

Actually, it's the other way around. Everyone has been fed up with ICANN because they've been dragging their feet on getting this done.

Re:Just Like Oil

[explosivejared]

The Universal Service Fund is evidence enough for you. Billions of dollars of subsidies wasted as windfalls to stockholders. Your lack of faith is wise, and it's only being supported by the new broadband plan laid out by the president.

It would be nice to have a perfectly efficient method of coercion to force ISP's to actually spend their subsidies on broadband penetration, but no one in power seems to be interested. It's the same story as IPv6 up to now. ICANN seems to be taking the lead finally. Hopefully someone will follow suit in the broadband arena.

Re:Just Like Oil

[joeytmann]

Even so, my orginal statement still holds true...everyone was sitting around waiting to see who would be the first to implent IPv6. Maybe I over assumed ICANN was the one that got fed up at someone else dragging their feet. For all practical purposes everyone was pointing fingers at someone else in a weird tug of war. It really doesn't matter who adopts it first, obviously with out DNS names associated to an IPv6 address its going to be a pain to get to a server, but just as obvious with out IPv6 addressing and routes what good are IPv6 DNS names? So now ICANN supports IPv6 DNS, now its upto the backbone hosts to get IPv6 implemented across the internet's core routers.

Re:Just Like Oil

[vertinox]

The only problem is people don't get very motivated until we're really on the edge.

The key problem with that is that the longer you wait the more equipment you have to replace and/or upgrade. And more upgrades and replacements means more money spent that would otherwise not had the equipment been installed with IPv6 from the get go.

So if you wait too long you can cripple your company/agency with overhead in contractor and equipment costs which could have been averted if you implemented a plan earlier. Of course, in the process you'll also make a lot of hourly contractors happy for the over time ;)

Re:Just Like Oil

[rmstar]

Do you know how deep people will drill for oil? Do you know how many kilograms of battery you need to substitute a kilogram of gasoline? No? Thought so :-)

We are not addicted to oil just because we are lazy. We are addicted to oil because it is so god-damn good. We will be badly screwed if it runs out, and no amount of innovation will bring such a wonderfully convenient energy source back. In comparison, and, come to think of it, not even in comparison, IP6 is a complete and total triviality.

Re:Just Like Oil

[Wesley Felter]

People are adopting IPv6 from the outside in, but they're using 6to4 and Teredo instead of the obsolete 6bone.

Re:Just Like Oil

[jguthrie]

The 6bone is only obsolete because they pulled the plug on it. My 6bone connection (to Sprint) was SOLID, even after I disconnected my direct feed to Sprint, but my tunnel broker connections have bounced up and down like rubber balls. That means that IPv6 was useful for me back in the days of the 6bone, but is now useless for just about anything. I don't know about 6to4 because I have never understood what 6to4 was and every new explanation of 6to4 seems to make less sense than the one before.

You might consider the last part a warning about the lack of utility of 6to4. If I can't figure it out, and I've been using IPv6 for, wow, it's a decade now, then someone who has no clue (and such people make up the bulk of the people you're trying to get to use IPv6) is not going to have a chance.

Re:Just Like Oil

[Wesley Felter]

Most people don't understand anything about IP, yet they use the Internet just fine. If your OS or router enables 6to4 automatically then you don't need to know anything.

6to4 is pretty similar to configured tunnels, but it structures its IPv6 addresses in such a way that each endpoint can automatically discover the IPv4 address of the other endpoint. Thus 6to4 requires no configuration or state in the network.

Re:Just Like Oil

[nuzak]

Oil is already not all that convenient an energy source. It's a scarce commodity subject to fluctuations, that has to be brought in from far afield and refined. Wind, solar, hydro, geothermal, don't have quite the yield, but when you're sitting on top of those resources like, say, Iceland is, you're doing great, and if you're a big country that has all those resources, it suffices to say the lights aren't going out due to supply problems. And there's nuclear of course, though that does have a few problems besides the waste. Oil is however a wonderful raw material. Just about every damn thing you touch in a given day is made in part or improved on with oil. It's a damn shame we waste so much of it by burning it.

And yeah, the chorus of people screaming about how IPv4 isn't going to run out sound a whole lot like the people who think we have limitless oil -- and perhaps we do, but in both cases, it's going to be damned expensive to retrieve and distribute.

Re:Just Like Oil

[jguthrie]

But that's just wrong. People DO know something about IP, even if they don't know that they know. They know that they have to buy IP connectivity from a provider, that they can't get on the Internet without that connectivity, and they know how much that provider charges for that service. Contrast this with your explanation of 6to4.

You say that "6to4 requires no configuration or state in the network" so how does the network know that I'm using 6to4? Does it read my mind? Do I have to write a letter to that effect and then burn it so that the smoke carries my plea to the 6to4 gods that will do the magic to establish connectivity? What? What is different about a computer that is using 6to4 as opposed to one that merely has IPv4 connectivity? I'm not trying to be difficult, I'm just trying to understand.

I suppose what's really going on is that there is some router somewhere that knows that one side has IPv6 addresses and the other side gets a feed from some place that knows what to do with the packets, but that router has to be configured and, given the current dismal state of the IPv6 connectivity in the world, that means that the end user has to know how to do it. What that means that it breaks the "the end user doesn't understand anything about IP" model which you describe (and I generally agree with...assuming that by "understand" you mean "knows any of the technical details of") in your first paragraph. Either that, or you're counting on the ISP to set that up and, well, if they're getting the explanations I'm getting (to wit: "there's no setup involved") then they're going to throw their hands up in the air and go play "Guitar Hero" at that point.

In other words, 6to4 appears to simply fail the "requires no configuration" requirement on the part of people who consider themselves competent to know what that means. It's easier to understand tunnel brokers, but a tunnel to one of a handful of tunnel brokers that is not networked near to me and with an uncertain bandwidth aggregated among a potentially large number of users lacks a lot of appeal to an ISP and, well, the lack of access equipment speaks for itself.

Of course, it's far from clear that keeping the 6bone around would have had the effect of boosting IPv6 adoption, because the lack of adoption among end users (my stats show something like a dozen IPv6 hits from outside my household LAN in the last three years) has more to do with the lack of access equipment that backbone availability, but it still sticks in my craw that they declared "IPv6 available to everybody" before it was really available to anybody.

Re:Just Like Oil

[Twanfox]

Unless you buy equipment capable of both IPv4 and IPv6 and use that for new installations. Then you would still have to replace only the current equipment that cannot handle the IPv6 schema.

Re:Just Like Oil

[Anonymous Coward]

Just like how when we run out of oil, solutions will come along, when we run out of IP addresses, solutions will come along.

Except that you want to prepare for when that day comes (in oil and IPs), so that when it arrives you're ready as opposed to scrambling and in a panic. We're all rushing towards the future at 60 minutes / hour.

Re:Just Like Oil

[sweatyboatman]

except when we run out of IP addresses, there wont be any new globally accessible servers. when we run out of oil, it wont matter how many IP addresses we have, because the modern world will completely collapse.

Same thing with climate change, the human race has taken no decisive steps towards curbing greenhouse gases. Despite the dire warnings of leading scientists over many years, despite modeling the greenhouse buildup's future effects, and despite a pretty universal feeling that everything is getting pretty f-ed up. I get the sense that we'll start doing something to address the problem right about when the Atlantic Ocean floods over Manhattan island. Of course, millions or billions of people will already be dead at that point.

But eventually solutions will come along.

What ICANN is doing here is actually something incredible laudable. They are actually taking positive steps to address (no pun intended) a future problem in a systematic way. I am sure in 2011 when IPv4 is getting close to used up, there will be a couple people frantically switching over to IPv6, but because IPv6 will have been around for so long, most people will have adopted it long before it ever becomes a problem.

Re:Just Like Oil

[Wesley Felter]

You say that "6to4 requires no configuration or state in the network" so how does the network know that I'm using 6to4?

The network doesn't know anything about 6to4, because 6to4 encapsulates everything in normal IPv4 packets which the network treats like any other traffic. Packets from one 6to4 host to another go directly between those hosts (no triangular routing like with tunnel brokers) and packets between a 6to4 host and a "native" IPv6 host go through a stateless relay which is located using anycast to a well-know IP address. The only special configuration required on the relay is to enable 6to4 relaying; it doesn't need to know which hosts will use it.

Re:Just Like Oil

[drinkypoo]

Just like you don't understand the problem with using oil, you don't understand the problem with using IPv4. The problem with using oil isn't what you do when you run out, but what it does to us in the process. CO2 levels have never been this high, EVER... well, okay, maybe during our volcanic period. But not since life took off on this planet, even through cataclysm! IPv4 has led to poisonous technologies like NAT which harm the peer-to-peer nature of the internet. The internet was meant to be a place for sharing of information, but the information is now held in castles and guarded by copyright. The client-server model has thus become pervasive, with asymmetric download/upload rates and centralized content (like blog sites.)

Re:Just Like Oil

[WhiteWolf666]

Offtopic, but---

It simply doesn't follow that Co2 levels haven't ever been this high. That Co2 that we are generating; you know, from fossil fuels?

Where do you think it was before it became fossilized?

http://www.geocraft.com/WVFossils/PageMill_Images/image277.gif

For most of the current Cenozoic era, Co2 levels have been *higher* than they currently are. The *only* possible issue with "global warming" right now is whether or not the rapid rate of change in Co2 levels will be damaging, not the absolute level of Co2 in the atmosphere.

For example, during the Jurassic period, Co2 levels were at 1800 ppm. During the Cambrian period, Co2 levels were 5000 ppm. Currently, Co2 levels are at 378 ppm, and even if we burn ALL known sources of Fossil Fuels it is unlikely we will drive that above 900 ppm or so.

Re:Just Like Oil

[rs79]

Don't forget the "Intellectual Infrastructure Fund" that NSF staffer Don Mitchell tacked on to domain prices when he told his contractor Network Solutions to begin charging for domains (much to their shock and chegrin; staffers at NSI would post checks to the wall and throw darts at them).

Don was a bit believer in the IETF process and the money was supposed to be used to keep that process alive - workshops, grants for airfare etc. People from all over the worls paid into it, Canada alone put $2M into it.

The US congress gave it to Mike Robers as a reward for going along with their secret plan to erect ICANN and scuttling the plans in place built up over the previous year by the community for the new company to administer domains.

Google for the wrapup meeting of the Boston Working Group to see th shock and horror as this unfolded.

Re:Just Like Oil

[rs79]

[quote]What ICANN is doing here is actually something incredible laudable.[/quote]

Um, yeah. Never mind it took the 8 years to do this. And that the V6 addy for F *changed*.

Never confuse "laudable" for "competant".

Re:Just Like Oil

[sweatyboatman]

Never confuse "laudable" for "competant".

fair enough.

Re:Just Like Oil

[Anonymous Coward]

turn landfill into oil?

Oh Noes The Internets!

[Apple Acolyte]

We're running out of tubes!!!

Seriously, though, I have a feeling that IPv4 will be saved by an ingenious tech solution far in advance of the world running out of addresses.

Re:Oh Noes The Internets!

[Anonymous Coward]

You mean like IPv6?

Re:Oh Noes The Internets!

[betterunixthanunix]

An ingenious solution like...IPv6? We have the solution already. We just need to get people off their behinds and get it implemented.

Re:Oh Noes The Internets!

[Anonymous Coward]

there is a solution, its called IP v6 :P

Re:Oh Noes The Internets!

[Clete2]

Yeah, just NAT everything! :(

Personally, I enjoy having my own IP address.

Re:Oh Noes The Internets!

[jez9999]

Seriously, though, I have a feeling that IPv4 will be saved by an ingenious tech solution far in advance of the world running out of addresses.

Yeah, there are loads of great solutions that you can just invent really easily.

For example, we could put the whole world on 1 IPv4 IP. Then we simply use a secondary metric for determining which host we want to send data to; this could be a 128 bit number, meaning there will easily be enough secondary numbers for all hosts in the world for a long time to come.

Re:Oh Noes The Internets!

[thwack328]

we could put the whole world on 1 IPv4 IP

And when the 4,294,967,297th world wants in? Then what???

Re:Oh Noes The Internets!

[timeOday]

That's what NAT already does; the secondary metric is the port number.

Re:Oh Noes The Internets!

[budgenator]

yes sir 64K should be enough for anybody!

Re:Oh Noes The Internets!

[timeOday]

Well, it's 64K * 4 billion. The 4 billion does help some.

Re:Oh Noes The Internets!

[budgenator]

What I was joking about is each TCP/IP stack only has 64K ports so the NAT box can only support 64K - WKS, Well Known Services; for a home network that's plenty, for a business with a hundred IP devices it'll still probably work, but an enterprise with a thousand or two IP devices there may be times when they just plain run out of ports on the NAT box. We got companies with printers exposed to replace faxes, VOIP telephony, video-conferencing, web browsers, instant messaging, databases, custom applications, Digital radiography, digital dental impressions, CAD/CAM mills, 3D "printers"; plenty of stuff trying to talk over the internet that didn't exist 5 years ago.

My home network allows over 10M hosts

[davidwr]

Sadly, it can't Talk dirEctly to my Next-DOor neighbor, who runs an equally large neTwork.

Re:My home network allows over 10M hosts

[Wolvie MkM]

Well played sir.. I wonder how many people that will woosh over...

Re:My home network allows over 10M hosts

[Anonymous Coward]

at least one
*rasies hand
tendot? huh?

Re:My home network allows over 10M hosts

[Ankur Dave]

10.0.0.0/8 is one of the IP blocks allocated as for private networks (ten dot star), but if hosts are in a private network they can't contact other private networks directly.

Re:My home network allows over 10M hosts

[Junta]

Unfortunately, it Looks like my network is more Awesome (over a septillion hosts, and less than one in a trillion chance it would conflict with my neighbor's)

Re:My home network allows over 10M hosts

[aztektum]

I was confused too. I didn't expect proper grammar so I added the S. Google has no clue wtf stendot is either.

Re:My home network allows over 10M hosts

[Blakey Rat]

Maybe he meant STENDEC: http://en.wikipedia.org/wiki/Star_Dust_(aeroplane)

Why punish them?

[austin987]

Why punish those companies for investing in (and giving a boost to) the early internet? If IPV4 is that important and ISPs need that many more addresses, it wouldn't be difficult for some ISPs to purchase blocks of IPs from those companies. Let the market do its thing. Besides, my NetBSD toaster is still waiting for a public IP...

Great news

[Besna]

As we keep hearing in other circles, "change" is in. As a software engineer who doesn't want to slow down for IT support, I have to do my own wrestling with the network. The more I can focus on the real job, the more productive I'll be.

2011

[Anonymous Crowhead]

It will be the year of IPv6, the Semantic Web, Perl 6 and rocket cars. It will also be the year of Linux on the Desktop.

Re:2011

[Anonymous Coward]

Not only that, but it will also be both ridiculous and not even funny.

Re:2011

[Tony Hoyle]

We've been hearing this 'addresses will run out by year x' for 20 years, and the predicted date has been wrong every single time. It's very hard to get enthusiastic about something that seems to be run by chicken little... Sure they'll run out eventually, and there's a network there to deal with it when it happens.. until then... zzzzzzz

If google, microsoft, redhat, CNN and the BBC (insert favourite site here) all go ipv6 (and by that I mean google starts indexing it too), that will be the year of ipv6. No way in hell it's going to happen before that.. I know of exactly zero useful ipv6 websites - I'm connected here but it's never been used.

Without any websites to actually *visit* on ipv6 ordinary users aren't going to go through the hassle, so ISPs see no demand and won't implement it (even though it would be a nice revenue stream for them - $10/month for 256 ipv6 addresses for example (and I really can't see them giving any more, seriously.. It's more likely to be 8 or 16 to separate the 'home' ($10/mo) users from the 'business' ($50/mo) users who get 256)).

Of course without any home routers that support it it's all moot anyway (hacked linkysys routers don't count).

Re:2011

[Anonymous Coward]

We've been hearing this 'addresses will run out by year x' for 20 years, and the predicted date has been wrong every single time.

Citation needed. I've heard many estimates, but none has ever been before 2010.

Without any websites to actually *visit* on ipv6 ordinary users aren't going to go through the hassle, so ISPs see no demand and won't implement it

Not this again. The technology that got us here is not the same technology that's going to move us forward! Web browsing, as you note, is not the killer app for IPv6. Maybe it's VOIP, or video, or P2P, or something we haven't even seen yet, but web will get dragged into IPv6 like everything else, not leading the way. That doesn't mean IPv6 isn't going to happen soon.

Of course without any home routers that support it it's all moot anyway (hacked linkysys routers don't count).

How about stock Apple routers?

Re:2011

[rs79]

For historical completeness I'd be remiss if I didn't point out there was a day when the "million name COM zone" was a fear so seemingly legitimate Postel himseld was gravely concerned. It wae felt the servrs wuould simpy fall over.

Thre are about 60M com names presently.

Re:2011

[linuxpyro]

Yeah, it's really too bad that the world's ending in 2012.

Good.

[Anonymous Coward]

Maybe we can finally get rid of the abomination that is NAT. Then all those Windows machines will become exposed to the Real Internet, and the false sense of security granted by their little home router will shatter the illusion of Windows stability once and for all.

Either that, or router manufacturers will start including SPI firewalls that aren't completely useless.

Re:Good.

[Anonymous Coward]

You're cute when you're ranting from your soapbox. Can I fuck you?

Re:Good.

[Macrat]

There is an illusion of Windows stability?

temporary fix

[JeanBaptiste]

just switch to IPv5 until things get sorted out

Re:temporary fix

[Ash-Fox]

just switch to IPv5 until things get sorted out

The Internet Stream Protocol (IPv5) doesn't redefine the way addresses are handled and thus offer no benefits over IPv4 in this particular issue.

IPV6, a lame solution for no problem

[Anonymous Coward]

So just because people waste IPv4 addresses by not using NAT and not recycling unused addresses, we want to force everyone to go to a solution that won't work correctly on existing devices that don't support v6, has a completely silly address, makes people get out from behind the elegant and awesome solution of NATs, and is basically poorly conceived, designed and executed?

Forcing v6 will be a disaster. It's better to force people to better implement v4 and take that time to design a system that will expand the address space while not causing so many issues.

This will be anonymous coward because I know almost everyone on /. LOVES change for the sake of change and anything shiny and new MUST be awesome, therefore I'll be modded down as an idiot and a troll for telling the goddamn truth.

Re:IPV6, a lame solution for no problem

[Anonymous Coward]

What truth? That you never worked with networks before? That you ARE an idiot? NAT elegant? My dick is more elegant than NAT.

Re:IPV6, a lame solution for no problem

[gaggle]

The elegant and awesome *what*?

You, sir, are trolling. Please explain yourself otherwise... I can't think of a single positive thing to say about NAT (other than the security aspect, but a firewall does that part of the job even more elegantly).

Re:IPV6, a lame solution for no problem

[Anonymous Coward]

The thing is that IPv6 is shiny and new (well not TOO shiny and new, because it has been in discussion since 1994 or so...) but it has not been battle tested, like IPv4 has. IPv4 has been through the growing pains of what happens when people try things that bend/break the standard, from unexpectedly large ping packets, teardrops, SYN flooding, and many other types of ways to bring machines down without requiring brute force in bandwidth.

IPv6 has yet to see daily sustained attack by people who want to find the next "ping of death" exploit to either cause machines to crash, or even better to find some way of finding a popular OS that a malformed packet could cause a buffer overrun, allowing code to be injected even if the machine is told to deny all incoming packets sent to it. People would easily pay millions if not more if they could find an exploit like this, and a chink in IPv6's armor could mean billions of dollars lost in downtime and security breaches.

Yes, IPv4 is old and creaky, but its been around for so long that pretty much all the show-stopper bugs have been found and ironed out, both in practice and in theory. IPv6 is still very young and even though it has been around in theory for about a decade now, it still hasn't been battle-hardened on a daily basis on production high-profile sites.

Re:IPV6, a lame solution for no problem

[Alioth]

NAT is an ugly hack. It's like a telephone extension lead that only lets you make outgoing calls.

Dynamic IP addresses suck too, for that matter: it's like a phone that changes number every time you use it.

more the story

[trybywrench]

The only justification you ever hear for moving to IPv6 is address exhaustion in IPv4. There's a lot of other stuff built into the protocol that will make the net a much better place. Even if IPv4 had the same amount of addresses as IPv6 it would still be worthwhile to switch. Just give this a once over for an introduction

http://en.wikipedia.org/wiki/Ipv6#Features_and_differences_from_IPv4

Re:more the story

[gclef]

Yes, lets take a look at those:

Larger address space
This is the address exhaustion argument.

Stateless address autoconfiguration (SLAAC)
Interesting, but not a selling point for users, and will make administrative management a pain in the ass. Most networks will use DHCPv6 to have records of which host had a given IP address...but they'll still have to run AutoConf to get a default gateway. This kind of split is annoying more than it's helpful.

Multicast
This is really only used on the link level, with one or two site-level things. I don't think this will not be used heavily. Also, if you want multicast, it's already available in IPv4. So this isn't really a gain with IPv6.

Link-local addresses
End users don't care, most sites won't care. In fact, the only people who do care are the authors of EIGRPv6 and OSPFv6 implementation. This isn't really a gain...just a difference.

Jumbograms The first possibly interesting thing in the list. It won't be used by many places, but DB->App server jumbograms are a common thing in IPv4, and making those bigger & standard is a reasonable gain.

Network-layer security
aka IPSec. Implemented, but key exchange is left as an exercise for the reader. (In other words, it's not happening.) This will be used very, very rarely. This is also something that's already available in IPv4, so not a gain for IPv6.

Mobility
Interesting, and also something definitely new....but not actually implemented anywhere. Not clear if this will fly at all.

No more checksum at the network layer
I'm not sure if anyone really cares.

In short, the single biggest selling point for the vast majority of businesses and users really is the extra size. The other stuff is either already available in IPv4, or only useful for some rare cases. In the majority of cases, the extra IP space is IPv6's only real selling point.

Re:more the story

[Kastigador]

Okay, okay, so I get the benefits. I just don't see this as being a very logical and easy to manage protocol. Perhaps I've just cut my teeth on IPv4 and shudder to think of having to adjust to completely new standard, but I seriously can't see myself ever looking at an IPv6 address and going "oh, that's just our gateway. or... oh no, that's somebody in Romania." It also truly scares me to even think that all IP's will now be routable. As if containing/restricting offending network traffic wasn't tough enough for a intermediate level network engineer. Forgive me if I resist, but there also seems to be a lack of "here's what the world will be like as a system's admin with IPv6" info out there. I forsee this being a long ways off, regardless of whether we run out of IPv4 address and prices get jacked up on everyone.

I don't expect much to change

[Waffle Iron]

ISPs see the limited IPv4 address space as a revenue stream. Many of them charge almost double for the privilege of getting a fixed public IP address. They don't have to spend money on a lot of scarce IP addresses themselves since they can always stick their customers in NAT ghettos.

They're not going to be very eager to give up their position as a gatekeeper of a limited resource just so their customers can frolick in a vast address space for free. Since most of them operate in a monopoly or duopoly situation, the proverbial "free market" won't force them to move off IPv4 either.

Re:I don't expect much to change

[Tony Hoyle]

Exactly... Expect 'cheap' accounts to be allocated within a 10.x.x.x net long before an ISP thinks of implementing ipv6. They'll probably pitch it as a security feature ('let us control the firewall for you! Surf in safety! Only $10/month!').

If a user wants a public IP. That's more cost. If they want a *fixed* IP.. go talk to the business services manager over there.

If they do implement ipv6 it'll be done the same way. 1 ipv6 address per account (ipv6 NAT exists and has done for a while). If you want 8 of them that's more cost. If you want more than 256.. see that guy in a suit waving? Go hand him your chequebook.

And before anyone says 'but but we'll all get 16 million addresses!'.. yeah, over the rotting corpses of every major ISP in the world.

Re:I don't expect much to change

[Kjella]

While I'm sure that's true, surely many ISPs must be running into the same problem themselves? I mean it's not like they have a magic bag of IP addresses to take from either, even with NAT. When enough of them get tired of trying to squeeze IPs out of other ISPs or enough pissed customers that don't get real IPs, they'll probably band togather and move to IPv6. I'm not holding my breath though...

Re:I don't expect much to change

[Diagoras of Melos]

It's not just the ISP's deriving revenue from fixed IPv4 addresses. Aside from all the corporate Class A's mentioned up top, there are hundreds or thousands of Class B's, and many of them list the possession of these address pools as an asset on their balance sheets. They are fought over in bankruptcy court. It's outrageous.

I used to work for Ampex, the inventors of the VCR, once a company with about 20,000 employees, now essentially a patent licensing firm with fewer than a hundred. They have a Class B: 136.185.0.0/16. That's right, more Class C address pools than employees.

There are lots of sunset companies in that situation.

Re:I don't expect much to change

[Danathar]

I expect once the address space runs out in IPv4 and new biz are FORCED to use IPv6 for their Web servers on the internet that there will be some screaming that public ISP's must provide v6 service to ALL customers.

If I want to start a business and it HAS to be on a v6 block because there are no v4 blocks left it would be a bit unfair competitively for me if the consumer ISP's refused to allow access to my web site due to lack of providing IPv6 to consumers.

I expect something like an edict from the Fed to command that (at least in the U.S.) that IPv6 be available to anybody who asks for it at some point.

Re:I don't expect much to change

[Anonymous Coward]

If they do implement ipv6 it'll be done the same way. 1 ipv6 address per account (ipv6 NAT exists and has done for a while). If you want 8 of them that's more cost. If you want more than 256.. see that guy in a suit waving? Go hand him your chequebook. If you consider /64 as a "single ipv6 address" then maybe.. otherwise you're on crack.

Re:I don't expect much to change

[Anonymous Coward]

And before anyone says 'but but we'll all get 16 million addresses!'.. yeah, over the rotting corpses of every major ISP in the world.

So.... What are we waiting for.

Re:I don't expect much to change

[glitch23]

And before anyone says 'but but we'll all get 16 million addresses!'.. yeah, over the rotting corpses of every major ISP in the world.

It's more than that though. The whole method of assigning IPs is a formal process. To get a certain block you have to be a carrier or something alone those lines. Just like with IPv4 addresses, any individual can not acquire an IPv6 address whether they want to or not. It just isn't going to happen. IP addresses aren't bought, they are leased. Last year the DoD paid for their pool which included a few billion IPs for lots of millions of dollars). No corporation owns their IPs. An individual would in theory go to their ISP as the next level in the hierarchy in an attempt to get their own IP. If an ISP is in the business of splitting up their IPv6 pool to individual users then you have a shot but it will cost money because they will have to spend money to use IPv6. If you attempt to bypass the ISP then you run into the constraints of the formal process preventing individuals from ever coming close to getting their own. There is the private block of IPv6 addresses that users are free to use just like IPv4 addresses from RFC1918. I believe the block is fc00/8 but I could be wrong. I believe Apple's OSX uses the block to assign an IPv6 address to network interfaces.

Re:I don't expect much to change

[Anonymous Coward]

"yeah, over the rotting corpses of every major ISP in the world."

Ideally, yes.

I'm guessing no more than 10

[davidwr]

Which is 1 less than 11 which is 1 less than 100.

Difficulties

[Wowsers]

Look, IPv6 is all well and good, but apart from typing 1:: for localhost, how am I going remember my outside IP?

Oh, and the line "There's no place like 1::" just don't sound right. http://www.thinkgeek.com/tshirts/generic/5d6a/

Re:Difficulties

[Midnight Thunder]

Look, IPv6 is all well and good, but apart from typing 1:: for localhost, how am I going remember my outside IP?

There shouldn't be an inside or outside for IPv6, since there is no notion of NAT. If you mean your router, then using a service like dyndns.org is an alternative.

Re:Difficulties

[Abcd1234]

Look, IPv6 is all well and good, but apart from typing 1:: for localhost, how am I going remember my outside IP?

Uhh, using the DNS name you registered?

Re:Difficulties

[kellyb9]

Look, IPv6 is all well and good, but apart from typing 1:: for localhost, how am I going remember my outside IP? Ummmm... right it down?

Re:Difficulties

[Dagger2]

Look, IPv6 is all well and good, but apart from typing 1:: for localhost, how am I going remember my outside IP?

Easy.

My IPv6 address is, shall we say, 2002:725b:3294::1. "725b:3294" is just my IPv4 address, which is apparently easy to remember. The 2002 is well-known, like the 192.168 in a private class C IPv4 address, and the ::1 is as hard to remember as the ".0.1" at the end of such an address.

In other words, it's no harder to remember than your current external+internal IP combination is.

Now, of course, if you deliberately pick an address like 2002:725b:3294:08d3:1319:8a2e:0370:7334, then yeah, you're gonna have trouble remembering it. But you can hardly complain about it if you chose it, can you?

Re:Difficulties

[Anonymous Coward]

What if I left it at home?

Re:Difficulties

[Atti K.]

Wasn't that ::1?

Re:Difficulties

[Alioth]

Fortunately, there's this brand new technology called DNS that makes the computer do it for you :-)

Re:Difficulties

[Kredal]

Then SSH into your home server, using the IPv6 address you... are.. looking...

oh, I see the problem. Hrm. Can't help you there. Sorry!

IP Squatting?

[Anonymous Coward]

Its the latest internet growth industry, IP Address squatting. For a slight, yearly fee I'll rent some of my IP addresses to you. :-)

There's a huge amount of IPv4 addresses still left

[bytesex]

If only Apple and IBM and stuff were to give back some of their A network space.

I have a solution

[obarel]

Why don't we use the 32 bit source address in each IP packet as an extension of the destination address? That way we can get 64 bit destination address.

Oh, wait...

Peak IP?

[misleb]

Have we reached Peak IP?

Re:Peak IP?

[Surt]

Depends on your interpretation. It is likely we have now created all the IP we are ever going to create. Allocation, however, is likely to continue to rise for many years to come.

Re:Peak IP?

[SJ2000]

Doubt it, RIAA is still kicking.

Blame it on the man!

[Zaphod The 42nd]

This is all Al Gore's doing, I just know it. Go make the internet again, and this time do it right!

Re:Blame it on the man!

[The Angry Mick]

Go make the internet again, and this time do it right!

It wasn't his idea to put a refrigerator on the 'net.

Re:Blame it on the man!

[Zaphod The 42nd]

Are. You. Serious? Holy crap, LG. I have no words. Oh wait, now I do; I can't wait for the day when I can hack into someone's LAN and change all their temperatures, ever so slightly. They'll never see it coming!

IPv6 migration behind a NAT

[Midnight Thunder]

IPv6 is not quite there yet, with some of the popular web sites still not accessible via IPv6.

If you are stuck behind a home router, with NAT then you will probably find yourself unable to access IPv6 sites. In the meantime there are two solutions:
    - Teredo. If you have Vista this is standard. For everything else there is Miredo
    - Aiccu. A litte more work and bureaucracy to get up an running, but a solution non-the less

Of course there is also Apple's Airport Extreme, which is one of the few home routers out there that support IPv6. I believe some of the third-party firmwares will do this too, but I don't think the IPv6 support is mature. As for Linksys, D-Link, et al. I think you are out of luck for the moment.

Also, if you running Apache, you will need a minium of Apache 2 and specify IPv6 support, using the configure script, prior to building it.

It's a sham - the Internet is mostly dark

[mcrbids]

The sad part is, most of the IP addresses in question are... dark. Nothing there. Even though we're approaching 85% allocation, utilization is probably around 1-2%. No, I'm not kidding.

Try it yourself - hack up some script to randomly generate IPs and then ping sweep the network blocks. You'll probably be quite surprised at the result.

A while back, I wanted to have a way to detect if a host was "offline" so that it could modify its behavior. (EG: halt outgoing SOAP requests if the server's network connection was disrupted, preventing bogus error messages from entering the system)

My first thought was to randomly generate 10 IP addresses, then ping them to see if they were offline, guessing that at least 50% would respond. Basically, none did. So, then I tried randomizing addresses and keeping a list of only those that had, at one time, responded. Even that turned out to be unfruitful. So finally, I took a dictionary and randomly created domain names from 1-2 normal dictionary words, pinging those, and keeping a list. That yielded some 40% usable responses, allowing me to keep a list of fairly trustworthy ping hosts to determine the online status of the server in question.

Bottom line: The shortage in the global IP pool is an artifact brought on by grossly inefficient/incompetent management of the global IP pool. The idea that we're running out of addresses purely ignores the fact that the vast, vast majority of the addresses we now have are simply unused.

Re:It's a sham - the Internet is mostly dark

[RedHat Rocky]

Ping?

Most large server farms block ICMP/ping at the border. Relying on ping to indicate whether an IP is occupied is just wrong.

Granted, I'm with you on the "large empty pool" theory.

Re:It's a sham - the Internet is mostly dark

[iowannaski]

Just because I didn't respond to your stupid ping doesn't mean I'm not using my IP address.

Re:It's a sham - the Internet is mostly dark

[plague3106]

Try it yourself - hack up some script to randomly generate IPs and then ping sweep the network blocks. You'll probably be quite surprised at the result.

Um, my router which has a public IP doesn't respond to pings. I think its set that way by default. I wouldn't be suprised if there were many IPs that ignore pings.

Re:It's a sham - the Internet is mostly dark

[dAzED1]

just because nothing responds, doesn't mean nothing is there.

iptables -P INPUT DROP
iptables -A INPUT -i lo -j ACCEPT
iptables -A INPUT -i eth0 -m state --state ESTABLISHED,RELATED -j ACCEPT

Tada. "reject" is one thing, because it actually responds. "drop" though - just drops the packet...almost like nothing is even there. And yet, I can go out and do things, and stuff can get back to me on established sockets.

Re:It's a sham - the Internet is mostly dark

[Aqualung812]

It isn't that simple. With IPv4, you have to subnet and that takes a few addresses. You'll have to have a gateway and a firewall/NAT device, there is two addresses. Network address and broadcast address take up two more, so now assuming you only need 1 IP address and took the smallest block, you just "wasted" three.

Now, maybe day one I only need 4 IP addresses. I get a subnet that can handle that, plus maybe 2 more. Now, when I need to add 2 more, I have to add a whole new subnet, waste more IPs, AND my ISP is going to have to get new ones. After all, they can't just have an unused subnet laying around, or else it would be wasted as well, right?

Getting back to end-to-end networks is what needs to happen (no more NAT), and IPv6 is the way.

Re:It's a sham - the Internet is mostly dark

[Rhys]

You're going to hit a lot of firewalls, many of them at the ISP level, that block your ping or traceroutes from going through. I wouldn't count on that as reliable.

Re:It's a sham - the Internet is mostly dark

[Hognoxious]

I once heard that MIT has enough IP adresses for everyone who's ever worked or studies there and will for the next 100 years to each have a unique one.

Re:It's a sham - the Internet is mostly dark

[Anonymous Coward]

Why would you ping random hosts? That's retarded.

Just ping the upstream router or gateway you moran.

Re:It's a sham - the Internet is mostly dark

[BitZtream]

While I would love to agree with you completely as I believe ARIN is a bunch of tards (can't speak for the other registries). There are/were technical reasons behind the way IPs are assigned. Machines haven't always had 2 gigs of ram. Maintaining routing tables on a network the size of the Internet was a difficult task, which required aggregating networks at upstream links and all sorts of stuff in a desperate attempt to prevent every multihomed router on the Internet from needing a few gigs to hows the paths to various subnets and determine what path was the best.

Of course, time goes on, ram is cheap, and doing it now is somewhat easier, but it still requires ram and processing power, and that increases latency and cpu utilization.

For instance, assume that everyone was assigned address space in blocks of 256 address (class C) and had to show they utilized the address space before getting more as well as prove they continued to use it. Now assume that only half of the address space available was assigned. 2.1 billion addresses in use. Thats approximately 8.3 million class C blocks
allocated. I'm going to assume thats higher than what we have actually in use these days (not allocated, in use) but bear with me for reference purposes.

Now, for each packet you route, you have to search through those allocated blocks and find the one that contains the address you're communicating with. You also have to determine which path of the many you may have on your router is the best path to use based on number of hops to the destination (we'll pretend AS hops are real hops for simplicity), include other factors such as your internal weights for a route because its expensive for you to use the OC3 you have rather than the DS3 because you got a great deal on the DS3 but not so much on the OC3.

You've just spent a lot of CPU cycles trying ot figure out which path to use. Now ... do this on hardware from 10-15 years ago. Well, first off, unless your at a NAP 10 years ago, doing this would require expensive memory upgrades on your routers because most didn't have the ram required to deal with a such a routing table in the first place, now add in the processing increase your going to need because even though you can cache routes and deal with updating the cache only as the external paths change, it only helps so much because those external paths change a lot so your cache hits have to be revalidated more often than you think. God forbid you have a flapping connection, as I can tell you from personal experience, on many routers from 15 years ago, a flap of a line that relays BGP information resulted in a router that was busy for a few seconds dealing with the BGP changes unless it was a fairly high end router.

So ... the point to all that is, a lot of the way address space was assigned was because the hardware we had to work with 'back in the day' was only capable of so much.

Okay, so now we can do better, great! Lets readdress everyone ...

I'm not going to bother going into the complexities of re-addressing a large network, but its rather a pain in the arse and can cost a whole hell of a lot of money in IT resources. So when you look at the big picture and think, 'well, I can readdress now and help deal with the problem and then have to eventually switch to the new protocol (for now, IPv6) eventually anyway OR I can wait till everyone has to switch to the new protocol because of this problem and only do it once'

It makes more sense to wait and do it at once, save yourself some money, deal with it when everyone else does, and deal with the least amount of work you can until that time. And ... this is how businesses make money, but not doing extra work they are just going to have to do again later if they can prevent it.

Of course, on that same note, there are plenty of businesses which don't exist yet that will make a killing off the scare of running out of IPv4 address space and the switch to IPv6 ... just like all the ones who made out over y2k fears/bugs.

Re:It's a sham - the Internet is mostly dark

[mxs]

You are an exceptionally bad engineer, coder, thinker, and internet citizen.

The sad part is, most of the IP addresses in question are... dark. Nothing there. Even though we're approaching 85% allocation, utilization is probably around 1-2%. No, I'm not kidding.

And you have ANY hard data to back that up ? No. Others are trying to come up with better metrics (http://www.potaroo.net/tools/ipv4/index.html is exceptionally verbose), but you ? You are not kidding about thinking that it maybe probably is around 1-2% ... Wow.

Try it yourself - hack up some script to randomly generate IPs and then ping sweep the network blocks. You'll probably be quite surprised at the result.

Bzzzt. No, I would not be -- nor should anybody be. First of all, it's not a requirement for every address to be routable to (and you can check that much better by looking at what percentage of prefixes are actually advertized). Second, many, MANY hosts and networks are behind firewalls, intrusion detection & response systems, etc. -- a "simple pingscan" can easily land you in a black hole at the network border after a couple of pings -- if access to those machines is even allowed from your network. Sure, in consumer broadband connections you don't often have such firewalls restricting inbound access, but that's not the "entire internet". Hell, go ping amazon.com and see what you get back. Nada, that's what.

A while back, I wanted to have a way to detect if a host was "offline" so that it could modify its behavior. (EG: halt outgoing SOAP requests if the server's network connection was disrupted, preventing bogus error messages from entering the system)

A problem many others have faced and solved before you.

My first thought was to randomly generate 10 IP addresses, then ping them to see if they were offline, guessing that at least 50% would respond.

Accounting for the different classes of addresses, unroutable space, bogons, etc. in that random calculation would be more work than the result is worth, especially seeing as how the state of netblocks can change over time. I wonder, why was your first thought to crap out (at least) 10 packets to the net that really are not needed ? What possible reason could there be for you to automatically ping a cellphone in Singapore ? Just imagine everybody doing this, just to check whether they are "online" ... How about choosing some well-known addresses (such as one of your own servers in a different locale, or possibly "well-known" servers that you know will respond and that don't mind a ping from you every now and then ... Not only do you get a 100% response rate when everything is working correctly, you also forego abusing bandwidth in remote locales you are not at all interested in.

Basically, none did. So, then I tried randomizing addresses and keeping a list of only those that had, at one time, responded. Even that turned out to be unfruitful.

You know, while still a bit dickish, it might have occured to you that most of {a-m}.root-servers.net do reply to ping or DNS requests. So do, in all likelihood, a router in your upstream, or DNS resolvers you know about. Instead, you now latch on to addresses that respond. The cellphone in Singapore, for instance.

So finally, I took a dictionary and randomly created domain names from 1-2 normal dictionary words, pinging those, and keeping a list.

Ah. So now that flooding ICMP out to the net is not enough, you have to litter it with bogus DNS requests the reply to which you are not really interested in. Again, imagine EVERYBODY doing this. Why not pick 10 known domain names and always ping those ? At least the results will be cached, and you may even choose ones whose owners you know and can ask whether they mind to be flooded with icmp every now and then.

That yielded some 40% usable responses, allowing me to keep a list of fairly

Re:It's a sham - the Internet is mostly dark

[trjonescp]

Try it yourself - hack up some script to randomly generate IPs and then ping sweep the network blocks. You'll probably be quite surprised at the result.

Dude, it's just corporate firewalls dropping ICMP packets.

Re:It's a sham - the Internet is mostly dark

[Ashtead]

Maybe 1 or 2 percent is a bit on the low side, but there is a lot of attrition with sub-netting /30s, where there are 4 addresses taken, of which two (one half) are the network and the broadcast address (x00 and x11) leaving the remaining two (x01 and x10) for actual devices, of which one will most likely be a router or gateway, and the last one is the actual server doing something interesting.

On the other hand, several ISPs use dynamic allocation when assigning IP-addresses to their customers, and a number of these addresses may not be in use at any one time, but as ADSL customers and dial-up hosts in particular come and go all the addresses are used eventually.

As for ping for checking, not everything wants to reply to pings anymore, for whatever reason: the fact that pings to some address don't show replies could just as well be because the host at that address is eating the packages, as compared to there not being any host there. An alternative which seems to be better able to tell the difference between a dead host and one that doesn't want to talk, would be nmap with TCP connection attempts at a number of likely ports, but that would tend to draw attention.

Re:It's a sham - the Internet is mostly dark

[Relic of the Future]

"The sad part is, most of the IP addresses in question are... dark. Nothing there. Even though we're approaching 85% allocation, utilization is probably around 1-2%. No, I'm not kidding."

You mea like this?

Re:It's a sham - the Internet is mostly dark

[gmack]

You can go out and do things as long as those things don't happen to be on a link that happens to have a smaller packet size than yours. Blocking all ICMP is a common firewall mistake.

Re:It's a sham - the Internet is mostly dark

[kestasjk]

In addition to routing IPv6 has so many addresses that we'll be able to give any device an IP address, so mobile phone networks will be able to be a proper part of the internet

Re:It's a sham - the Internet is mostly dark

[TemporalBeing]

Getting back to end-to-end networks is what needs to happen (no more NAT), and IPv6 is the way.

That's assuming I want all my devices to be publicly visible. What if I don't? While NAT is a little PITA to set up, it works beautifully for the job. I don't want people to be able to easily figure out the all the systems on my network, and even if I converted my network to IPv6, I want a solution like NAT.

NAT just makes it easy for the network to have a single point-of-contact going in/out of the network.

And Firewall issues would still be the same - as far as having to poke-holes, etc. And not-having firewalls would make for a rather in-secure network and not solve any of the problems that we have today any way.

So the issue really is an IP allocation issue, and NATing would be good regardless of using IPv4 or IPv6. It would be nice for everyone to be able to have a static IP at their network gateway, but not beyond that.

Re:It's a sham - the Internet is mostly dark

[jhol13]

What if I don't? Just firewall it.

While NAT is a little PITA to set up, it works beautifully for the job. Just like firewall.

I don't want people to be able to easily figure out the all the systems on my network, and even if I converted my network to IPv6, I want a solution like NAT. I think you still can use NAT, if you really, really want. But a rule "block everything from outside, except established connections to to machines in subnet X and http to web server" is as easy as same in NAT box.

It would be nice for everyone to be able to have a static IP at their network gateway, but not beyond that. Unless, of course, VoIP is considered.

Re:It's a sham - the Internet is mostly dark

[jhol13]

ping sweep the network blocks. You'll probably be quite surprised at the result. ~$ ping microsoft.com
PING microsoft.com (207.46.197.32) 56(84) bytes of data.

--- microsoft.com ping statistics ---
5 packets transmitted, 0 received, 100% packet loss, time 4002ms

YES! Microsoft is dead!

Re:It's a sham - the Internet is mostly dark

[mcrbids]

Your response is littered with lots of inflammatory words which work rather effectively at reducing the value of the substance in your post. Despite some initial hesitation, I'll bite...

And you have ANY hard data to back that up ? No.

Handily ignoring the rest of my post?

Others are trying to come up with better metrics (http://www.potaroo.net/tools/ipv4/index.html...

Which gives lots of details about ALLOCATION. See rest of post.

First of all, it's not a requirement for every address to be routable to (and you can check that much better by looking at what percentage of prefixes are actually advertized).

Eh, what did you say? And, why use an address that's not advertized? (spelled "advertised" by most folks) Sounds amazingly like NAT... in other words, it is not used in the literal sense.

A problem many others have faced and solved before you.

It's like asking "Do you know where my shoe is?" and hearing your spouse respond: "Right where you left it". Technically true, utterly useless in any meaningful sense.

I wonder, why was your first thought to crap out (at least) 10 packets to the net that really are not needed ?

Ignoring the insults, my intention was to try to evaluate the relative strength of my Internet connection, including an accounting for outages that can cover broad areas. By picking addresses from around the world, I'm attempting to evaluate the overall strength of the Internet rather than just my connection to a few key points.

Ah. So now that flooding ICMP out to the net is not enough, you have to litter it with bogus DNS requests the reply to which you are not really interested in.

Are you daft?

You're worried about the effect of about 10-20 DNS requests per day, when mail servers are pelted with millions of SPAMs per day? Have you ever looked at the relative size of a ping packet (a few bytes) with the size of an image-laden penis-pill spam?

Come back when your sense of scale approximates reality.

So, do, in all likelihood, a router in your upstream, or DNS resolvers you know about. Instead, you now latch on to addresses that respond. The cellphone in Singapore, for instance.

I wasn't interested in my personal connection to my upstream, but rather a more general sense of my host's health. And how many cell phones actually *need* a public IP address, or even have one? I argue that an "unadvertised" address is a waste of an address, function better served by NAT.

The fact is, if no changes in allocation and revocation of IP space are made, there is a cutoff date when no more IP addresses will be available.

A point which underscores my post, even if you disagree with my methods.

Re:It's a sham - the Internet is mostly dark

[MyIS]

Small nitpick - about individual cellphones not needing public IP addresses. What about VoiP? Cellphones, too, already rely on public address entries - plain old phone numbers. One way or another, there are billions of existing uses of public addresses, except that they are all kludged into separate custom namespaces. And it is already necessary to bridge them in expensive and complicated ways.

Re:It's a sham - the Internet is mostly dark

[mxs]

Your response is littered with lots of inflammatory words which work rather effectively at reducing the value of the substance in your post. Despite some initial hesitation, I'll bite...

You may find them inflammatory, I find them warranted given the evidence.

And you have ANY hard data to back that up ? No.

Handily ignoring the rest of my post?

Which does not provide any useful data, just hypothesis after hypothesis.

Others are trying to come up with better metrics (http://www.potaroo.net/tools/ipv4/index.html...

Which gives lots of details about ALLOCATION. See rest of post.

Ah, you completely missed how they are also looking at advertized space ? Also, fractions of allocated space that are also advertized ? I argue that that is a BETTER metric than what you came up with by far.

Eh, what did you say? And, why use an address that's not advertized? (spelled "advertised" by most folks) Sounds amazingly like NAT... in other words, it is not used in the literal sense.

It's quite possible that I misuse advertized vs. advertised, it's been a while since I got graded on English tests (and no, it's not my native tongue). Why use an address that's not advertised ? Consider the case where two large and different networks have to interconnect. RFC1918 addresses will clash there, easily. Also consider addresses that are limited in scope -- for instance military networks. There are valid reasons. Not all reasons for this are valid, but neither are all of them a sham.

A problem many others have faced and solved before you.

It's like asking "Do you know where my shoe is?" and hearing your spouse respond: "Right where you left it". Technically true, utterly useless in any meaningful sense.

I should have added the word "better", you are quite right.

I wonder, why was your first thought to crap out (at least) 10 packets to the net that really are not needed ?

Ignoring the insults,

Insults ? I'm not attacking you, I'm attacking your method. You are littering the net.

my intention was to try to evaluate the relative strength of my Internet connection, including an accounting for outages that can cover broad areas. By picking addresses from around the world, I'm attempting to evaluate the overall strength of the Internet rather than just my connection to a few key points.

So you don't actually "just" want to see whether your host is "up" and "connected", but rather want a poor man's internet weather report ? (http://www.internettrafficreport.com/main.htm and such ?) Your approach doesn't even accurately the relative strength of your internet connection, as you well noted (hard to predict how many of the hostnames and ips you come up with SHOULD respond; when you actually pick them carefully (for various characteristics, such as networks you want to monitor connectivity to), you actually have a gauge of how well your connection works compared to what you would expect. Furthermore, if you save the hosts that have responded in the past, there is a very real chance that you get an unrepresentative hostlist that won't accurately gauge connectivity problems "on the wide internet". Even that, however, can easily be done by using well-known addresses and network paths instead of randomly shooting stuff at addresses you are probably not interested in at all.

Are you daft?

I might very well be, but not on this matter.

You're worried about the effect of about 10-20 DNS requests per day, when mail servers are pelted with millions of SPAMs per day? Have you ever looked at the relative size of a ping packet (a few bytes) with the size of an image-laden penis-pill spam?

You may be surprised to discover that I have. I'm not worried about a single person doing it. I'm worried that a whole lot of brai

Re:It's a sham - the Internet is mostly dark

[dAzED1]

so? What is your point?

Try it yourself - hack up some script to randomly generate IPs and then ping sweep the network blocks. You'll probably be quite surprised at the result.

The claim that non-responsive IPs mean they're not being used is absolutely silly. I'm not going to argue with you if blocking incoming untrusted ICMP is a "mistake," but it is certainly very common. Which is, well, my point. What was yours?

IP6 won't matter til Google supports it

[wowbagger]

Wake me up when I can pull up the main page of Google using nothing but packets with IP6 headers.

That means that I can do a DNS query using nothing but IP6 packets - NOT IP4 packets.
That means that I can do an HTTP transfer from Google's servers using nothing but IP6 packets - NOT IP4 packets.

Hell, wake me up when there's a AAAA record for Slashdot.

This is a *baby* step towards IP6 being useful.

Re:IP6 won't matter til Google supports it

[gbjbaanb]

This is a *baby* step towards IP6 being useful. Yup. The thing with first steps is that you have to do them in order to make the second step (obviously), but then you can make a third and fourth steps. Next thing you know, you've got to where you were going.

Now Google can register an AAAA record, do you think they will? If they couldn't register one, do you think they would?

Re:IP6 won't matter til Google supports it

[Anonymous Coward]

That sounds just like somebody saying "the web won't matter til Microsoft supports it" in 1995.

If the giant in the field started actively supporting it, that would be one way for it to take off. Another way would be for somebody else with a killer app to use it, stealing Google's fire.

In fact, that's how a lot of these big technology transitions occur. It would be pretty surprising if Google managed to weather a few of them as top dog -- that's rather rare. We can see they've got advertising, search, and big databases down. Can they transition to the Next Big Thing? I'll believe it when I see it.

Re:IP6 won't matter til Google supports it

[Cajal]

This is actually a very important step towards what you want. About two-thirds of the TLDs have authoritative servers which are reachable over IPv6. There's a complete list at my blog - http://www.personal.psu.edu/dvm105/blogs/ipv6/2008/01/ipv6-dns.html

So you can query the root and .com DNS servers using IPv6. If you want Google to be reachable over IPv6, go talk to Google. Everything higher in the tree is IPv6-enabled now. And Google has an IPv6 allocation from ARIN - they got a /32 2005 - http://ws.arin.net/whois/?queryinput=!%20NET6-2001-4860-1

I agree that there isn't much content on the IPv6 internet now. So if you want it, yell at the content providers.

Consumer router support

[nsayer]

I've been waiting a while for Netgear, Linksys and that crowd to add 6to4 support to their home NAT routers as a way to help jump start IPv6 adoption. There would be no security issue if incoming connections were blocked by default and people could turn it off if they didn't want it. But 6to4 can be set up automatically by any machine with a publicly routable IPv4 address.

Well, I'm happy to say that my wait is finally over. They didn't make a big deal about it, so I don't know exactly when they did it, but Apple added that support to their Airport Extreme. So now when I go anywhere that has one of those, I can directly SSH into those inside machines that I've opened ports for without undue muss or fuss.

Apple has been a stalwart supporter of IPv6, from my observation. It's been possible to use AFP file sharing over IPv6 since at least Tiger and the built-in VNC stuff works over IPv6 too (though there is a naming lookup bug that requires you to connect using the IPv6 address literal if you use the command-K "Connect to" dialog).

So, Netgear and Linksys, what's holding you guys up?

Re:Consumer router support

[saleenS281]

Netgear: your guess is as good as mine.

Linksys: It's called cisco. They'd be happy to offer you 6to4 capabilities, but that's a *business* need, so you'll need to pay for their *business* class hardware.

Re:Consumer router support

[nsayer]

Linksys: It's called cisco. They'd be happy to offer you 6to4 capabilities, but that's a *business* need, so you'll need to pay for their *business* class hardware. I guess building hardware to keep up with the no-default routing table for IPv4 is too lucrative.

Of course that's the wrong attitude. It's not a business need. Moving to IPv6 is a basic survival strategy for the Internet. If Cisco doesn't want to invest in helping to bring the transition closer, then pressure should be brought to bear to show them the error of their ways.

Re:Consumer router support

[bar-agent]

If Cisco doesn't want to invest in helping to bring the transition closer, then pressure should be brought to bear to show them the error of their ways.

Yeah, that's gonna happen. Q: "What does an 800-pound gorilla do?" A: "Whatever it wants."

Cisco would probably say that, given their market dominance, their ways are not erroneous.

I get a surprising number of IPv6 hits...

[Omnifarious]

I get a surprising number of IPv6 hits on my webserver at home. Most of these appear to be XP or Vista boxes with Internet connection sharing turned on that automatically assign themselves a 6to4 addresses when they have an interface with a public IPv4 address.

IPv6 with 6to4 is easy to set up, and I'd recommend it to anybody who has a static IPv4 address. You can use NAT-PT so all your IPv6 hosts can still get to the IPv4 network. If you have a couple of DNS servers, you can even set up reverse DNS for your IPv6 network just the way you want using this nice web interface from the NRO.

I maintain some good links to stuff about IPv6 on del.icio.us.

I hate NAT. And I think IPv6 can be just as secure. Partly because a 64-bit address space is really hard to effectively randomly probe working addresses and partly because it's fairly easy to configure a firewall to not allow incoming connections.

Re:I get a surprising number of IPv6 hits...

[Jugalator]

Speaking of security, IPv6 also has IPsec part of the standard, which IPv4 doesn't.

Actually, IPv6 has a large amount of features besides the increased address space.

http://en.wikipedia.org/wiki/IPv6

Before IPV6 gets popular, it needs:

[JoeD]

Before IPV6 gets popular, it needs:

1. Home routers that support it.
And/or
2. DSL and cable modems that support it.

I'd love to convert my home network to IPV6, but as long as I connect through an IPV4 ISP, and my wireless router only does IPV4, I'm hosed.

Joe D

Re:Before IPV6 gets popular, it needs:

[nsayer]

Before IPV6 gets popular, it needs:

1. Home routers that support it.

Done.

Re:Before IPV6 gets popular, it needs:

[Atti K.]

Done, even cheaper ;)

Re:Before IPV6 gets popular, it needs:

[nsayer]

even cheaper ;) Linux folks always say that.

I guess it's true if your time has no value.

Re:Before IPV6 gets popular, it needs:

[Atti K.]

Of course it has. But the difference between this and this (or your favorite router) is much more than the value of my 1-2 hours (max) to set the thing up for IPv6. I'm just not paid that well :)

What about NATs

[llZENll]

So when IPv6 finally does become the norm, will there be any need for NATs on home routers, or will ISPs simply give you many addresses?

Re:What about NATs

[Bigon]

NAT are an evil hack to the ipv4 address exhaustion.. I think that ICANN recommends to the ISP to give a /48 pool to the end user, hope they will follow the recommendation (or at least give a /56)

Re:What about NATs

[Just Some Guy]

So when IPv6 finally does become the norm, will there be any need for NATs on home routers, or will ISPs simply give you many addresses?

Given a standard end-user allocation of a /64 network, you will have 1.8*10^19 addresses to play with. It is unlikely that you would need to fake more with NAT.

Re:What about NATs

[imemyself]

Yeah, but I really have my doubts as to whether consumer ISP's will actually give their customers anything more address wise than they have now. Just because its recommended doesn't mean that ISP's will stop being dicks about it.

Re:What about NATs

[Todd Knarr]

They'd have to really work at it. The normal IPv6 link configuration process only allows the ISP to set the leading /64 prefix. The trailing 64 bits are entirely determined by the host, and can't be forced to any particular value by the upstream router. It's possible to force the issue by way of MAC registration and DHCPv6 or a firewall, but it takes a fair amount of configuration to do it and it doesn't work well with the Windows IPv6 stack (which much prefers that you use stateless autoconfiguration of interfaces).

Re:What about NATs

[knorthern knight]

> Given a standard end-user allocation of a /64 network, you will have 1.8*10^19
> addresses to play with. It is unlikely that you would need to fake more with NAT.

That is really F***ing stupid...

- back in the old days, nobody imagined how much demand there would be, so they handed out Class A (16 million addresses) at a time. Then they discovered that addresses weren't really infinite, so they ended up having to invent CIDR and NAT to conserve addresses.

- mark my words; down the road, thanks to these idiotic IPV6 allocations, there will have to be a CIDR-like hack for IPV6 to reclaim some of the throwaway bits. WTF do individual users need *THAT* much space? A /104 would give every user the equivalant of a current Class A to play with. Heck, a /120 allows for 256 machines at my residence. I don't have that many electrical plugs, and the electrical bills would bankrupt me.

Re:What about NATs

[Just Some Guy]

No. They picked the length they did for routing purposes. So many bits indicate the top-tier ISP. Another set indicate the second tier. The rest are basically "bonus" that are given to the customers. In other words, 64 bits would be sufficient to manage the routing tables to give an address to everyone, and the extra 64 are what gives everyone the ability to ping their hairdryer.

The point of those 64 end-user bits is autoconfiguration. You are entitled to hand-number your machines like we do now, but there's no need to. Think of it as a 64 bit hash rather than a 64 bit integer.

BTW, CIDR is basically impossible in IPv6 (thankfully).

Re:Great, IPv6, an insecure protocol

[Just Some Guy]

Lest anyone think this jackass is correct:

IPv6 barely supports firewalls or NATs, allowing any Joe Sixpack to see what your secured corporate network topology is like from anywhere.

It is not up to the protocol to support the hardware. And anyway, all good firewalls support IPv6 already. NAT? It's there if you're dumb enough to want it.

It also does not support reserved IP blocks... change ISPs, and you are forced to re-ip your whole network.

Step one: update your router to the new netblock.

Step two: sed -i'' 's/^old:net:block/new:addr:ess/' db.mydomain.com; rndc reload

Step three: laugh at people who go around changing ISPs all the time.

Of course, IPv6 has -zero- hooks for IP level encryption, so this has to be handled at the trensport or app level.

If only it support IPSec, "the goal of [which] is to provide various security services for traffic at the IP layer, in both the IPv4 and IPv6 environments." Oh, wait...

NAT Sucks

[JSBiff]

NAT is, well, better than nothing, which, currently, is your alternative. But I'd hardly call it an "elegant and awesome solution". IMO, ultimately, NAT sucks because you *do not have a globally routable address* for devices in your network. Sure, that gives some security benefits, but makes it a PITA when you do want to open connections directly to a computer or consumer electronic device in your network.

A few reasons you might want to have a public address inside your network:

* Direct VOIP telephony (SIP, Skype, various instant messenger clients, run a TeamSpeak Server), etc

* Running game servers, web server, mail server, etc

* Remote access (VNC, SSH, etc)

* Direct file transfer with a friend (I've, from time to time, run into problems with things like instant messenger client based file transfers not working behind a NAT - though they do seem to have somewhat alleviated that problem - I suspect by routing my file transfer through the IM network instead of directly to the other person), or P2P file sharing systems, like Bittorrent - yes, they can usually work behind NATs; but they work better if direct connections could be more easily made).

Yes, yes, I know about port forwarding. That's fine and dandy as long as you only have a single device per port that you want to allow incoming traffic to. Ultimately, IPv6 is a much better solution to the problem of address space limitations than is NAT. NAT usually requires software to do ugly hacks to get around the limitations of only allowing outbound connections. A simple firewall with every device having a global address is a better solution, because then I can open up as many ports to as many devices as I like, without having to worry about only allowing one device per port.

I've had a number of times where I've been extremely frustrated by NAT. Often times, if software isn't explicitly written with NAT in mind, and the problems it creates, then it won't work well in a NAT'ed network.

Re:NAT Sucks

[Anonymous Coward]

NAT has a number of advantages though:

1: People see one external box, and have to crack that box to get to your internal network segments. Yes, this can be regarded as security through obscurity, but this keeps someone who is "driving by" with some autodiscovery tool from gleaning info they shouldn't have.

2: An attacker has to figure out if the box with a web server is one machine, or actually multiple, with the router redirecting ports. For example, if there is an attack that requires something done with both the FTP server and a SSL server at the same time, it won't succeed. Another example is having the SSL port to one machine and the non-secure Web server point somewhere different, or having the dynamic Web stuff hanging off of a different port than the static (which is not all good -- a lot of businesses block Web stuff that isn't going to port 80.)

3: A NAT box allows one to protect traffic, and deal with an abuse problem internally rather than have an outside person come in. For example, if someone is sending out obnoxious content, without a NAT, the outside place can bypass the net admins and try legal action against the owner of the machine. With a NAT, they would have to go through the company or organization's security (including legal team).

4: Legal reasons. If someone is being prosecuted for hacking, it gives a better case to show that the knowledge of internal network segments is protected and shielded, forcing the defendant to bypass security.

5: Business intelligence. Its always good to keep the number of machines (and what segments they are on) hidden, so the competition can't easily find out that one is ramping up a new backend infrastructure for a service rollout (for example.) This also goes for foreign intelligence as well. For example, if country A finds out that country B is adding a lot more computers to their IPv6 segment of a certain type in their infrastructure, it can bring meaningful info that country B may be ramping up for a military offensive.

6: Contracts. In a lot of security contracts, internal traffic and external Internet traffic have to be completely separate (separate IP address space), or else severe criminal and civil penalties can ensue.

7: Corporate laws like SOX, HIPAA, and PCI compliance. These laws make NAT a requirement. Fail to do this as a network or security admin, and you just lost the "due diligence" protection. This can mean shareholder lawsuits and prison time should a security breach occur.

Yes, NAT is ugly, but its something that is a must have on the Internet for most companies, even with the vastly larger address space of IPV6. NAT is also the law in a number of countries (as a consequence of "due diligence"), and not protecting internal assets by this could mean civil and criminal liabilities.

Re:NAT Sucks

[vux984]

A simple firewall with every device having a global address is a better solution, because then I can open up as many ports to as many devices as I like, without having to worry about only allowing one device per port.

Right, but then we need our routers to have that simple firewall. I want to be able to managed the firewall in one place, not on each device.

I also like that my LAN isn't dependant on my ISP being up. ie I don't want to rely on my ISPs DHCP and DNS. If my internet connection goes down, I want my lan to keep working.

So if we get rid of NAT, fine, I'd love to have a block of routable addresses instead. But I -still- want to manage a firewall at the router, and perform dhcp, and dns on the LAN, possibly even from the router.)

Unfortunately I can't name a consumer priced product that can do this for ipv4, never mind ipv6.

Re:NAT Sucks

[DaleGlass]

You don't need a firewall on each device.

Under NAT, you have one box, which has a table that says "port 80 goes to 192.168.1.3", "port 25 goes to 192.168.1.7", etc.

Under IPv6 with a firewall, you can have one box, which has a table that says "Connections from anywhere to ab:cd:ef::01 on port 80 are accepted", "Connections from anywhere to ab:cd:ef::02 on port 80 are accepted", etc.

I don't get where people get this strange idea that while NAT may control multiple computers behind it, a firewall somehow couldn't. The "protection" NAT provides is that it's a firewall with a "deny by default" policy, which is trivial to do without NAT if you want it.

Re:NAT Sucks

[vux984]

You don't need a firewall on each device.

I know that.

Under NAT, you have one box, which has a table that says "port 80 goes to 192.168.1.3", "port 25 goes to 192.168.1.7", etc.

I've used NAT, but thanks.

Under IPv6 with a firewall, you can have one box, which has a table that says "Connections from anywhere to ab:cd:ef::01 on port 80 are accepted", "Connections from anywhere to ab:cd:ef::02 on port 80 are accepted", etc.

Yes I understand how a firewall works too.

I don't get where people get this strange idea that while NAT may control multiple computers behind it, a firewall somehow couldn't. The "protection" NAT provides is that it's a firewall with a "deny by default" policy, which is trivial to do without NAT if you want it.

I have no misconceptions at all about this.

The issue has nothing to do with needing NAT. The issue is that we like the convenience of our NAT boxes, and there aren't any consumer priced firewalls that I'm aware of. I'd be delighted to use a sub $100 wireless router that had an honest to god firewall capability configurable via a friendly web client.

And without one, I need to set up a software firewall on each device. (which isn't a bad thing), but I want a single standalone box to manage network security. If I want to block port 25 inbound, I want to do it at the box, not on each unit.

I **KNOW** a firewall can do this. However, nobody sells one at the price point of a NAT box.

Care to name one?

Re:NAT Sucks

[DaleGlass]

The issue has nothing to do with needing NAT. The issue is that we like the convenience of our NAT boxes, and there aren't any consumer priced firewalls that I'm aware of. I'd be delighted to use a sub $100 wireless router that had an honest to god firewall capability configurable via a friendly web client.

Well, I use a Linux box for this. Get a cheap Pentium, attach a decent sized heatsink to it, use a CompactFlash to IDE adapter and a CF card, and you should be able to run that from a fanless power supply. With such low power usage you could probably even harmlessly disable the fan on a normal supply. The hardware can be found for free.

You can also get a Linksys WRT54G, those are below $100, and there's alternate firmware for it with plenty of functionality. Some googling suggests the WRT54G can be made work with IPv6 fine.

Re:NAT Sucks

[Anonymous Coward]

Its interesting that this issue is not brought up and discussed more. I understand your point of view about the kludginess of NATs, but as a architect for a large networking company, having worked with scores and scores of corporate networks all over the world, my experience is that NATs are standard in the corporate world, and whether that comes from ignorance or legitimate security concerns is an interesting discussion, but lets not forget the practical reality of the observation.

The fact of the matter is that, currently, NATs are here and they have to be dealt with. Protocols developed long ago, such as FTP, which used embedded IP addresses and separate control connections have been enormous challenges in the networking industry. I have written NAT proxies that support FTP properly, and I can tell you, it is a major pain to get it right, especially when you deal with thousands of concurrent connections churning through the port numbers. *sigh*. So, single connection protocols such as ssh/scp or http cause much less trouble and that is nicer for the networking folks to work with.

In the last decade, a lot of media protocols have become very popular, but unfortunately, in many cases the designers of these protocols simply ignored the issue of NATs. While it is fine to climb up to the top of the ivory tower and declare NATs are bad and your protocol should not be bothered with them, please do not be surprised to hear that hundreds of expensive networking software engineers in scores of different networking companies have to read a 70 page Masters Thesis to understand how to parse your protocol, and of course, then write and maintain tricky, mission critical network protocol software for years on end to deal with it properly. http://www.cs.columbia.edu/sip/drafts/Ther0005_SIP.pdf [columbia.edu]

Is it any wonder SIP has grown much slower than it should have given the underlying wonderful flexibility of the higher level semantics? But no, a budding internet phone service provider ends up buying and setting up complex and expensive SIP NAT traversal devices (google that phrase) just to get going. I suppose protocol researchers do not spend a lot of time working with corporate networks. Unfortunately, the cost of this ignorance has been enormous. Fortunately for most, it was swallowed by large networking companies who have not complained enough perhaps.

By the way, protocol researchers should look at SCTP as the basis for signaling protocols. It is based on IP and is an alternative to TCP and UDP see http://www.isoc.org/briefings/017/ [isoc.org] , and every operating system is on board ... except Microsoft, of course. But there is a standard 3rd party library available ala winsock. Perhaps, just as they hated winsock and the internet (and still do, IMHO) until it became too popular to ignore, perhaps so it will be with SCTP.

--
I do not want to write forever, but reviving the question from my first paragraph, it is interesting to think about whether corporations will be comfortable giving up the anonymity and security benefits (if only illusionary) of their NATs when they are presented with the opportunity to provide an unlimited number of cheap, routable IPv6 addresses to their employees. I honestly have no idea. It would make my job easier, but surfing slashdot with my personal corporate IP address would make it hard for me to be ... an Anonymous Coward

Re:NAT Sucks

[vux984]

Well, I use a Linux box for this.

A linux box gets the job done. But its just not a consumer friendly product on the same level as a NAT box. Not that they couldn't -make- one. But to my knowledge they haven't yet.

I wonder if it would be possible to get a hybrid ipv4NAT/ipv6Router+Firewall, where it'll route and firewall ipv6 at the same time as providing NAT on v4?

That would be perfect for the home market, provided we are even allowed to have blocks of ipv6 addresses.

Re:NAT Sucks

[JSBiff]

But here's the thing. The GGP says "we don't need IPv6" because NAT does everything *he* wants. You can still, I believe (I'm no IPv6 expert, and haven't played with it much, but if I get some free time it's something I want to play more with), do NAT with IPv6 if you really *want* to. My problem is that the attitude that "we don't need IPv6 because NAT solves the problem" is just wrong. It solves *some* problems, but not other problems.

Going to the points in your list:

1. Potentially, a Firewall, I think, could 'hide' the network behind it by blocking ICMP/Ping traffic as well as all TCP/UDP traffic that isn't explicitely allowed. I think a decent firewall could also implement similar techniques as NAT devices do to track when an incoming connection is related to an outbound connection (like for FTP where you have a 'command' socket, then when you transfer files, a data socket - though they have passive FTP which deals with that issue pretty nicely I think).

2. I'm not sure you couldn't do something like this without NAT - maybe by assigning multiple addresses to a single box? In any case, for situations where people really *need* NAT, it should still be possible with IPv6, I think?

3. Well, you could also use some sort of anonymizing proxy service, like Tor, perhaps. This particular point won't help most users, much. In the current IPv4 + NAT regime, a potential litigant could still track down the IPv4 address to the person who owns the NAT box, which is probably you, or your dad, roommate, etc.

4. Again, I think that to a large degree, a good firewall could accomplish much the same thing - again, the cracker has to penetrate the firewall (illegal) before accessing the internal network (except those ips/ports that are open on the firewall). I think that would still maintain a pretty high legal barrier.

5. God I hope my country's military is behind a super-good firewall. In fact, if computers are *that* sensitive, why are they on the Internet at all? And if other countries *aren't*? Well, what do I care - that's just to our advantage. *grin*. Seriously though, a good firewall should be able to hide your network fairly effectively, except to the extent that your network makes outbound connections which might be seen.

6. Setting up seperate, isolated subnets does not require NAT. That's not even an argument. Sure, you can use NAT for that. But, even with IPv4, there are like 2 seperate networks running through the same switches at my company. The switches and routers take care of isolating their traffic.

7. Ok, I'm no regulatory compliance expert, so I can't really speak to this, but why wouldn't a decently configured firewall accomplish the same "due dilligance"? It seems like *most* of the arguments people make for NAT are really arguments for *some kind* of firewall, but NAT is just one type of firewalling mechanism. But, again, I'll go back to the basic statement that IPv6 and NAT shouldn't be mutually exclusive.

Sorry, but I fail to see how *any* of these are an argument against IPv6, or that NAT isn't *the* solution for all problems. Some problems, maybe it is the best solution.

Re:NAT Sucks

[yorugua]

My problem is that the attitude that "we don't need IPv6 because NAT solves the problem" is just wrong.

What about the following sentence: 'My problem is that the attitude that "we don't need NAT because IPv6 solves the problem" is just wrong.'. I also find myself thinking about said sentence. When people comes to me with the "no more NAT because we have IPv6 and you can have an IP address for every cell of your body", I wonder if they really want me to, for examle, assign a publicly addressable IP address to every internal machine for example inside a bank. Is that a good idea? By my standards of today, I say no way. So, think about it: a) My problem is that the attitude that "we don't need IPv6 because NAT solves the problem" is just wrong. b) My problem is that the attitude that "we don't need NAT because IPv6 solves the problem" is just wrong. What's left? That is where IPv6 has to win, but for certain scenarios, NAT seems to be a "good thing (tm)".

Re:NAT Sucks

[xaxa]

My ADSL2+ modem/router/NAT/wireless-thing runs a Linux kernel, and uses something like iptables (I haven't looked very closely) for port-forwarding rules. There's a web interface to configure it. It was 'free' with my connection, which is pretty normal here (UK). It's not the only home router that runs Linux either. I don't think the consumer would notice anything different either, since it finds the hostnames of everything you've connected and the port forwarding is pretty simple to do, opening up a deny-by-default firewall should be the same.

Re:NAT Sucks

[TemporalBeing]

But here's the thing. The GGP says "we don't need IPv6" because NAT does everything *he* wants. You can still, I believe (I'm no IPv6 expert, and haven't played with it much, but if I get some free time it's something I want to play more with), do NAT with IPv6 if you really *want* to. My problem is that the attitude that "we don't need IPv6 because NAT solves the problem" is just wrong. It solves *some* problems, but not other problems.

I think I'm agreeing with you - but here's my basic opinion:

IPv6 should only be about expanding the address range. It is not a solution to give out IPv6 addresses to all devices and get rid of NAT; nor is it sufficient to say that NAT solves all problems and we don't need IPv6.

Reality is that IPv4 addresses will run out. So will IPv6 if we just let everything on the planet have an IP address - but that's not necessary. Reality is that we should convert our IPv4 addresses to IPv6 and pretty much leave the NATs and firewalls the same. We'll have to do that eventually any way.

Reality is that DHCP is a very effective method for handing out IP addresses, and the new "dynamic" IP assignment built-into IPv6 is (IMO) junk. A network administrator should be able to determine the IP address of all systems on the network. While dynamic addressing like what is built-into IPv6 is nice in theory, in practice it is not good; and any competent network admin or network security administrator will know why.

We really need to change how we are thinking of IPv6. For the most part, both sides are wrong; and until we can get that through, we won't really get the right solution.

Re:NAT Sucks

[lachlan76]

DHCP is available for IPv6, though.

Re:NAT Sucks

[LinuxDon]

Quote: "I **KNOW** a firewall can do this. However, nobody sells one at the price point of a NAT box."

That's because a firewall is pointless if you have to drop all incoming connections anyway because you're using NAT. The "firewalls" today are targeted at a somewhat niche market, causing them to be so expensive.
Believe me, as soon as these home routers support IPV6, they WILL include a firewall! (for the same price)

(The reason is that implementing a good NAT implementation is a lot harder than implementing a simple firewall.)

IBM may actually use a lot of 9

[Anonymous Coward]

I work for them (but obviously do not speak for them) and I personally have 8 machines with 9. IP addresses. Times that by a third of a million staff and add in whatever servers, managed services, infrastructure etc, you're getting to needing a class A.

Re:IBM may actually use a lot of 9

[Torvaun]

Do all of your machines need to be publicly accessible? Subnets for the win.

Re:IBM may actually use a lot of 9

[Wesley Felter]

Do all of your machines need to be publicly accessible? Subnets for the win.

Public machines use subnets too, you know. Besides, the promise of IP was one address for every machine, not every public server.

(Speaking as someone using many addresses inside 9/8.)

Re:IBM may actually use a lot of 9

[drinkypoo]

I wish I could meet the guy who modded you up and beat him with the RFCs. Then I'd come for you. TCP/IP is a peer to peer networking protocol. NAT is a hack, not a good thing. It's the least bad thing given the circumstances. Yes, EVERY system should have a unique IP. Every PDA, every computer, et cetera. It's your firewall's job to make those machines accessible or not.

Re:IBM may actually use a lot of 9

[TheRaven64]

Accessible is not the same as addressable. NAT allows machines to be publicly accessible but not publicly addressable. With IPv6, every machine is publicly addressable but not necessarily publicly accessible. The nice thing about this is that it means that a machine has the same address irrespective of where you are. If you have a machine at home, behind a NAT, that you SSH in to, it has an internal and an external address. You have to know which side of the NAT you are on to decide which one to use. With IPv6, the address will be the same for both, which simplifies this process. You may still only have the SSH port open to the outside world (or you may not have any ports open) but that has nothing to do with the address.

Re: Reply to: Peak IP?

[Anonymous Coward]

Lets find out!

can you ping me at fe80:806f:0:SomeShit123... oh just frack this crap... I'll FAX you the IP!

Projections

[mxs]

"Projections suggest that this unallocated pool will run out by 2011 at the latest.'"

Riight. Last I read it was 2011 for ARIN, 2012 for RIPE, assuming current allocation procedures. If allocation- and revocation-procedures are changed before then, "at the latest" suddenly becomes "at the earliest".

There is a problem, but it's not got a final due-by date attached to it just yet.

Re:Great, IPv6, an insecure protocol

[Anonymous Coward]

Interesting post... but totally irrelevant.

With no ability to NAT or firewall in IPV6 , anyone on the external Internet can find out exactly what you have, theb stage targeted attacks on every single host on a private network.

As for the stated IPsec, it was a nice draft... but never made in the standard.

IPV6 is a net admin's worst nighmare for security, next to unpatched machines.

Dear Slashdot User

[Anonymous Coward]

Hi, Jeff Immelt here. We here at GE own the 3.0.0.0/8 (and other misc. /16's and such). Basicly, as we're the 4th largest company in the world, you can suck it. We like to allocate willy-nilly and/or as we see fit these networks on our internal infrastructure. So what if it is rarely in a contiguous manner so that it is impossible to summarize in our routing tables. We got money, we'll buy bigger stuff.

Now that I just PWND your face I have to go see about maybe implementing this cool new technology called "vlans" I keep hearing about. IPV6. PSSSHHHH...maybe around 2040.

-Jeff

Double byte AS numbers are running out too

[MavEtJu]

Not only the IPv4 IP space is running on empty, at the last AusNOG conference (a must for everybody who is into internetworking) a talk was given about the similarities and differences in the allocation of AS numbers.

Where the IP space allocation graph shows an exponential line since 1990something, the AS number allocation graph shows a linear line.

The interesting thing is that somewhere in 2010/2011, when the IPv4 IP space is running out, also the double byte AS number allocation is running out. At around the same time!

So while the big world has to deal with the IPv6 (which by now should be common knowledge and practise), the ISP world has to deal with the four byte AS numbers.

That last part isn't 100% true: If you have a double byte AS number and your BGP speaker doesn't understand four byte ASN numbers, you will see some strange things in your BGP table but everything will keep working. On the other hand, if you have been handed out a four byte AS number, you'd better make sure you got a speaker which supports four byte AS numbers :-)

oh grow up

[Quadraginta]

I'm sorry, but this seems like a whole bunch of arrogant cluelessness rolled into one.

First of all, it certainly used to be easy to be an ISP, and there used to be a whole slew of mom 'n' pop ISP operations. The fact that a lot of those closed up shop has much more to do with the fact that they just couldn't compete with the cutthroat pricing the telcos (and to a much more limited extent) the cable companies were rolling out to acquire market share. If you're living in an area with only one or two ISP choices, it's very likely to be the fault of you and your peers in ranking low price above every other possible consideration. When customers do that, they drive everyone out of the business except the largest and most diversified company, which alone can turn a reliable profit by multiplying teensy-weensy margins on each account by zillions of accounts, or can even take a loss on providing network service because they make it up on portal ad revenue, selling add-on services, whatever.

If you think a market exists for ISP service with fixed IPv6 addresses, nothing's stopping you from opening up shop, hiring the necessary network engineers, and trying to cover their salaries, medical benefits, and 401k's with your subscription fees. And if you aren't willing to give that a spin, 'cause you lack expertise to even know WTF you're talking about, or you do actually realize that people generally won't pay for that kind of service, then you're being cynical and dishonest in complaining that other people won't go into the business and lose their life's savings trying to shovel back the tide for your temporary benefit.

Secondly, the reason ISPs charge more for fixed IP addresses is just because it makes some of their network administration more complicated. My ISP (Cox cable) used to give out fixed IP addresses ten years ago, but switched over to DHCP and variable addresses as their customer base grew, and presumably the complexity of administering fixed addresses when customers might add the service and then drop it a week later, and then add it again in three months (and demand their old address back) became incompatible with charging a price not too far above what Pacific Bell charged for DSL.

Finally, your complaint where it isn't uninformed seems to boil down to the complaint that people in the business of selling Internet access want to charge as much as possible for doing so. Uh...and so? Did you just fall of the turnip truck yesterday? Is it an outrageous surprise to you that people everywhere want to maximize the price they get for their labor? Don't you want to get the highest salary you can for the work you do? What makes the people who work for ISPs, or who start them, any different?

Re:oh grow up

[nuzak]

> If you're living in an area with only one or two ISP choices, it's very likely to be the fault of you and your peers in ranking low price above every other possible consideration.

Look, I'm as much a supporter of the little guy above the big impersonal corporation as the next, and I'm willing to shove some sheckels in the direction of that philosophy, but when the only differentiating factor you can determine is price, then price is going to be the decider. Water seeks low ground, markets seek low cost. Most people just want the internet as a utility, and simply either don't understand or don't care what the difference is.

If you don't differentiate your more expensive product, you die. If you can't differentiate it, you're selling a commodity, and you have to price accordingly.

Re:oh grow up

[Waffle Iron]

there used to be a whole slew of mom 'n' pop ISP operations.

Yeah, and they were on dialup. That doesn't count because nobody cares about dialup anymore. Broadband is a monopoly or duopoly due to the huge capital costs tied up in phone or cable lines, and 3rd parties ISPs have to outsource to the same telcos they compete with, which is a conflict of interest that doesn't make business sense. It has nothing to do with consumers choosing one feature or another.

Secondly, the reason ISPs charge more for fixed IP addresses is just because it makes some of their network administration more complicated.

Maybe a tiny bit. But the main reason they charge more is because the people who care about fixed addresses are almost certainly running some kind of server, and the ISP figures that they can extract more money from people who run servers. With IPv6, it wouldn't be nearly so easy to sort out who's running a server and who's not.

Finally, your complaint where it isn't uninformed seems to boil down to the complaint that people in the business of selling Internet access want to charge as much as possible for doing so.

You need to work on your reading comprehension. Where did I say that I was complaining? I was just explaining why you shouldn't be surprised that you'll probably be on IPv4 for a long time to come. Personally, I'd be happy not having to learn this whole new pile of alien IPv6 concepts.

You are right

[m0rtadelo]

The solution is clearly the WDS (WorldDHCPServer). One to rule them all.

Re:You are right

[DBCubix]

Is that going to be broken the British healthcare system, where I'll have to wait 2 years for an IP address? No thanks.

Re:You are right

[Anonymous Coward]

Is that going to be broken the British healthcare system, where I'll have to wait 2 years for an IP address? No thanks.

Mail it to 1999.

Quote from thisbbc news article

"In 2007, no one is waiting longer than six months for non-emergency treatment - down from between 18 months and two years for some surgery in the early days. By next year no one should wait longer than 18 weeks."

The six months figure only applies to a few cases; typical is 3 months (or less - can be a few weeks if you're flexible about dates/times/amount of notice).

(And no, I'm not a propagandist for the current Labour government; I'm probably voting Lib Dem at the next election, and I think the NHS dentistry coverage is scandalously poor.)

IPv6 -- Innovative Suicide

[Anonymous Coward]

The problem with the proponents of IPv6 is the rabid focus on eliminating NAT in the network. Rarely have I seen such an enthusiasm for cutting one's own throat.

Let's take a hypothetical example of IPv6 actually fulfilling the dreams of its proponents, such that NAT is removed, every device on the network has its own directly reachable IPv6 address, and there are indeed enough addresses that every household appliance can have its own address.

Congratulations. What do you do when IPv8 comes out? Or even IPX (sic)? When you have locked every device on the planet into a single network protocol with a dead-end to dead-end design, how do you ever manage to upgrade that to something better? Unless people want to claim that IPv6 is the ultimate perfection of network protocols and will never be surpassed, ever. Otherwise, you have to assume that some day, something better will come along. If people think the transition from IPv4 to IPv6 has been painfully slow, that's nothing compared to the idea of the transition time required for upgrading an IPv6 network to something else, when you would have to convince everyone to upgrade every single gadget they have with an IPv6 stack to the new technology? The concept of NAT and its associated technologies are a requirement to allow people to continue to develop new networking protocols and to allow transitions between technologies.

Without NAT, how do people plan on upgrading from a global IPv6 network to what comes next? The harder people make it, the more it slows down any future network protocol innovation.

Network protocols and applications should be designed to be NAT-friendly, not NAT-hostile. The easier it is to provide on-the-fly translation between protocols, the easier it is to deploy new technologies in networking protocols faster. Arguing against NAT is like arguing that people shouldn't use document converters because they break on MS Word documents, and everyone should just use MS Word documents, only, as a dead-end to dead-end solution. People should be striving to use protocols and applications that can be translated/converted cleanly between networks (yes, it's hard) to allow people to choose the technology/protocol of their own choice while still being able to interoperate with everyone else.

Re:IPv6 -- Innovative Suicide

[fbjon]

Since when does IPv6 disallow NAT? Use it if you want to!

Re:Great, IPv6, an insecure protocol

[Just Some Guy]

With no ability to NAT or firewall in IPV6 , anyone on the external Internet can find out exactly what you have, theb stage targeted attacks on every single host on a private network.

End-user netblocks are 2^64 addresses in size. If an attacker could ping a billion hosts per second, it would still take them 585 years to scan a single block.

So, again, NAT-as-security is even dumber on IPv6 than it is on IPv4.

As for the stated IPsec, it was a nice draft... but never made in the standard.

From Wikipedia:

IPsec is a mandatory part of IPv6 (mandatory to implement, not mandatory to use), and is optional for use with IPv4.

Wow. Guess you're wrong there, too.

wanna get sadder?

[keeboo]

Look at this:
003/8 May 94 General Electric Company

So GE has a whole /8 for them (3.x.x.x).

And now, look at this:
www.ge.com has address 216.74.131.56

NAT

[macdaddy]

Insert obligatory statements about why NAT is bad and why people who argue in favor of NAT don't know a damn thing about networking.

Re:Great, IPv6, an insecure protocol

[Anonymous Coward]

Step one: update your router to the new netblock.
Step two: sed -i'' 's/^old:net:block/new:addr:ess/' db.mydomain.com; rndc reload

Watch all your SSL certificates break, all your SSH configs break... hopefully one isn't using any software that is licensed to a specific address.

Step three: laugh at people who go around changing ISPs all the time.

Laugh at people stuck with being rate-raped by ISPs because the internal costs of managing all of their *internal* applications that depend on things like SSL and SSH are too high to allow them to easily change ISPs. If only they could have used an internal private address space that allowed them to isolate their own network from such changes, so they could always threaten to take their business else where if an ISP is not responsive...

Re:Great, IPv6, an insecure protocol

[growse]

I firewall ipv6 very nicely, thank you very much.

And you're last comment proves you're not a net admin.

USF

[macdaddy]

Not all ISPs abuse the USF like the RBOCs. I work for a small independent telco which covers a number of rural exchanges. We were offering DSL in towns of 200 people long before many of the larger metropolitan areas like Kansas City. We're currently deploying ADSL2+ and FTTH in areas to ramp up our bandwidth offering, provide triple-play and reach further into the farthest most confines out of exchanges to offer service to customers that previously had only POTS. All of this is being paid for by the USF in a cost-recovery manner. The town we CLEC with cable services are all ROI-based.

So, in short, not all ILECs abuse the USF. Some of us actually use it for the purposes in which it was intended.

Oh, and I'm working on having IPv6 deployed this year. Good luck finding more than a handful of residential ISPs offering IPv6 in this country.

Re:USF

[TooMuchToDo]

I'm working on putting togeter a co-op FTTH. I know you're part of a for-profit, but I wouldn't worry much, I'm out of the Chicago suburbs =) Any advice you can provide?

Re:USF

[macdaddy]

I wish Slashdot had private messaging capabilities.

Hit me offline: routerstud at the Google Mail domain dot com. (don't laugh; I didn't pick it...) We actually do consulting for service providers too.

Re:USF

[wowbagger]

IIRC you work for SKTC, right?

Re:USF

[macdaddy]

Hello, David. It's been a while. How are things?

Re:USF

[wowbagger]

Pretty good. I was about to leave SKTC to Pixius because of cost/speed, but then the line went to 1.5M and I lost my motivation to do so. You are correct - just because the RBOCs are bastards doesn't mean that all telcos are being irresponsible with the USF.

Re:USF

[macdaddy]

The speed issue is definitely something we're tackling right now. The previous DSL speeds were just atrocious. We have raised the speeds on all DSL customers in our Western areas. We can't do much for those in the Eastern areas with our existing infrastructure I'm afraid. We are replacing the entire DSL infrastructure over East with a distributed ADSL2+ solution to remedy the situation. In your neck of the woods I have good news. We will be deploying FTTH in our Western rural areas which should include your area (I don't know your exact location but I know that you're on DSL which we only deploy in the countryside). I don't have an exact timeframe for you but I expect it to be available by mid to late summer. The initial offering will be voice & data only. Video will come in a future project.

Hopefully that will give our users the "fat pipe" they need for the next generation of Internet technologies. The bandwidth estimates for online content like Netflix online movie downloads are unreal. I don't believe we've decided what speed options to offer but I imagine that it would at least be comparable to the cable speeds in our CLEC communites: 4, 7, 12.

Re:USF

[wowbagger]

Since this is venturing OT for this thread, I'll just ask this: Is there anybody I can contact to get on the list for this? I'd like to think I'd be a good test candidate....

email me, if you would be so kind, at wowbagger at (the company about which we are talking). net

Re:USF

[macdaddy]

I'll keep you on my list. So far we've only been doing internal testing. We have a lot of fiber work to do in the plant before external testing can proceed. I'll see if I can't get you included in a beta program when the plant is ready though. Have a good weekend.

Re:USF

[wowbagger]

OK, I'm in the Harvest Valley estates.

BTW: If you guys have a problem with pulling fiber the last mile, and want to look at going wireless - I have a tower, we can bargain.

Re:Great, IPv6, an insecure protocol

[Just Some Guy]

Watch all your SSL certificates break, all your SSH configs break...

Um, why? "myhost.example.com" won't change hostnames just because its address changes.

hopefully one isn't using any software that is licensed to a specific address.

Name one. Seriously. Name a package that is bound to a particular IP.

Routing derived from TUBA

[jd]

Basically, you can tell a huge amount from the routing address. You should be able to tell from the start of the address as to what type of communication it is. You then have two byte pairs, identifying the upstream gateway and the downstream gateway. Backbone routers must have a short prefix, followed by zeros or ::, followed by a suffix identifying the machine. The longer the prefix, the further from the top tier it is. There will therefore be a very well-defined set of prefixes that identify each country - one prefix for every upstream connection. Most nations will probably run top-tier routers, so the national prefix will necessarily be very short.

Mobile nodes will have the last 48 bits in their address equal to their MAC address, but ISPs that support mobility cannot pre-allocate anything in the valid range of MAC addresses. So, any IPv6 address whose ISP is known and whose suffix is arguably in the range for MAC addresses but is not a valid MAC address is connected to an ISP that does not support mobility. Likewise, if the ISP is know to support mobility, you can examine the last 48 bits to determine the nature and type of network device being used. Converseley, if multiple machines on that segment have resolvable MAC address style addresses, it's certainly autoconfiguring and probably suppoting mobility. Where routers fall into this category, the routers are autoconfigured and network mobility (NEMO) is probably supported. Network mobility is not useful except for failover (explained later) for fixed systems, so you're likely looking at a router that has a wireless connection upstream and must cross between two providers, so is most likely on a vehicle with moderate range and multiple on-board networked devices. An aircraft or a train is realistic, as is a military vehicle acting as the group's servive provider, but a car is unlikely (not enough machines inside), as are mass transit buses (they don't have the budget) and roller coasters (nobody's stupid enough to use wifi on one, and the entire circuit would probably be in range of a single router with a bunch of wireless access points to extend the range, if they were).

As one of the first IPv6 system admins, I can say that the biggest change is that you will need to learn the new syntax for using numerical addresses in a URL. It changes, because of a rather careless piece of symbol overloading. Actually, only an idiot would use the number if a name is present. Names work the same way. Adding in static IP addresses into BIND becomes a pain, so don't. Use a dynamic system, which means either DHCPv6 or IPv6' own automatic configuration and discovery. Then, you will never have to concern yourself with manually entering in IPv6 addresses backwards, one byte at a time, for no obvious reason besides someone having a sadistic moment.

For routing, the differences between RIPv2 and RIPng, or OSPFv3 and OSPF6, are very small. Most routes can be discovered, as they always have been. The routing protocols were developed before IPv6 had a default route - for a long time, it was believed you should have paths discovered as far as possible, so that you don't end up hard-coding single points of failure that don't exist in the network topology itself. (If some upstream router A fails, then traffic moves to upstream router B, as if the router was mobile. The NEMO protocol allows all upstream routers to re-adjust the paths to get to the router concerned efficiently without dropping too many packets.) Otherwise, current IPv6 routing is nearly identical to IPv4 routing.

For multicasting, service discovery is done via multicast, so it'd better be available over more than a local network, otherwise everyone is going to have to provide everything. Either that, or users will need to configure things for different networks. Which they aren't going to like. A VoIP cellphone should be capable of working over any wireless tower and get a service identical to the one they receive on their local service, though not necessarily at the same cost. IPv6 would not eliminate roaming, but it would eliminate the need to lose provider-specific capabilities when doing so.

Re:Great, IPv6, an insecure protocol

[Monkeybaister]

I can think of one: CheckPoint.

That's why any software that licensed to an IP address gets a private address from me. It's complete crap that companies think IP addresses couldn't change.

Re:Great, IPv6, an insecure protocol

[Anonymous Coward]

Um, why? "myhost.example.com" won't change hostnames just because its address changes.

My bad, I should have said IPSec (which I think you boast about in IPv6), which can use IP addresses.

Name one. Seriously. Name a package that is bound to a particular IP.

Uh, you've never used a license manager, have you? Let's see... hmm. I seem to recall HP Openview used IP address information in generating a license key (good heavens, can you imagine the fun it would be to have a network management tool running on a good sized network and then changing all the IP addresses on it at once?) I seem to recall Oracle also using such a license manager... oh, right! It's that little thing called FlexLM, node-locked licenses based on IP addresses... though to be fair it seems they may now support hostnames, too, sort of. I mostly remember the installations auto-generating license keys based on IP addresses, though.

Well that explains a lot

[PalmKiller]

I noticed the AAAA records when I did a dig the other yesterday to get updates of my root servers. I thought I needed to update something cause it looked off, while I realized it was ipv6 addresses, I said those shouldn't be there. As far as needing to get allocations, your current ipv4 allocations will work fine on the ipv6 network once its fully meshed will it not? After all all the ipv4 addresses are just grouped under one ipv6 allocation block...so why would they want people to rush out and get ipv6 allocations, or do they just mean new network allocations?

i try! i really do!

[Anonymous Coward]

look. I've been on the IPv6 wagon for as long as i can recall (so maybe thats
not that long ;-) ) - I've got the IPv6 on our nets, I've got our main servers
doing the v6. I've got the firewalls doing it etc etc but I still find
horrible broken apps and appliances around every damn month. I spend
most of my free time trying to submit patches to fix small utils
that have NO CLUE as to what a :::: address is! if we all get together
onto this one requirement... and get everyone who runs winXP to run 'ipv6 enable'
in their DoS box then we might start getting somewhere.

oh yeah. all those wireless captive portals. how many of those you reckon do
IPv6? :-(

Re:HEY TACO - LET'S GET WITH THE PROGRAM!!

[Hucko]

This is off topic, not a troll. It is a genuine problem! There should be a setting to load (n > 10) messages. I turned off the Discussion 2 system because it only allows me to load 50 post at a time. There are only a limited number of posts worth reading.

ICANN sucks

[Anonymous Coward]

I'm sure these bureaucrats have never seen an IP address.

It sure is going to be fun dealing with shit like 45345:34523587:1375:1209:4812aej:f3r:13298fh2:8h:f238hfwf:329:0fh13:98hf9:18h:f18hf:13:f8983hf.

Get all the myspace kids off the Internet along with all the other worthless trash and we'll have an abundance of IPv4 addresses.

VOTE RON PAUL - ONE HOUR LEFT [ronpaul2008.com]

Re:Great, IPv6, an insecure protocol

[Just Some Guy]

I admit that I was unaware there was any software that broken by design. Honestly, that's the kind of thing that needs to cause heartache to its users and authors to discourage more of it from being written.

Re:Great, IPv6, an insecure protocol

[dkf]

IPV6 is a net admin's worst nighmare for security, next to unpatched machines. Any admin's worst nightmare is "users". Without those pesky users, they'd be doing just fine!

U.S. Military required to be ipv6 by 2012

[dooguls]

http://www.ipv6.com/articles/military/Military-and-IPv6.htm
is just one example showing how the U.S. Military is required to be all ipv6 by 2012, in fact there's large chunks of the network that are supposed to be moving to IPv6 before then. So I'd say that's your "ball-rolling" starter. I have no idea how many networks and computers the U.S. Military represents, but considering they have an entire TLD, I assume they have a few. And I'd also be willing to bet that all the big router & OS vendors out there don't want to loose a big fat juicy customer like the U.S. Military, and therefore will do whatever it takes to get that network up and running.

You know in some senses, I think using the military as a guinea pig for things like this is a good thing for federal tax dollars to be spent on.

Re:Great, IPv6, an insecure protocol

[TemporalBeing]

With no ability to NAT or firewall in IPV6 , anyone on the external Internet can find out exactly what you have, theb stage targeted attacks on every single host on a private network.

End-user netblocks are 2^64 addresses in size. If an attacker could ping a billion hosts per second, it would still take them 585 years to scan a single block.

So, again, NAT-as-security is even dumber on IPv6 than it is on IPv4.

Wrong. It only takes one PING to be successful. Remember - security through obscurity (which is all that your suggestion would be) is not security at all. Use NAT and control your network better. It's not a dumb solution - it's smart security practice and helps limit the foot-print of a network on a larger network - also smart as it minimizes the points someone has to compromise and maximizes what they have to figure out/know to do so.

Re:Great, IPv6, an insecure protocol

[Just Some Guy]

Your main point seems to be that of controlling your network, and with that I wholeheartedly agree. I just think it's ironic that you're both advocating NAT and claiming that I'm for obscurity through obscurity. It's not inherently more secure than a proper stateful firewall.

Then there's the class B's

[rs79]

I know of one organization that has two class B's. They use about 150 addresses in one and none on the other. I know of another class B thas was bought for 100K frm a private party about 7 years ago. It's about 3% used.

I'll all for V6 addressing but will never use V6 addresses. There are other schemes besides V6 that work as long as hosts can process V6 addresses. They've worked for a couple of years now. So, hurry up and adopt V6 so we can bypass ARIN/RIPE etc.

127.0.0.1 doesn't have an 8-bit mask

[Laebshade]

Localhost (127.0.0.1) has a 32-bit subnet mask, so 127.0.0.1/32.

Re:127.0.0.1 doesn't have an 8-bit mask

[Xipher]

The loopback netblock is a /8, so even though your host may only use 127.0.0.1 itself, the entire 127.0.0.0/8 block is allocated for loopback addressing and is not usable on the public internet.

Re:127.0.0.1 doesn't have an 8-bit mask

[timbo234]

ocalhost (127.0.0.1) has a 32-bit subnet mask, so 127.0.0.1/32

It may be setup this way on your computer's network settings but the RFC says the whole /8 is valid is part of the loopback:
http://www.faqs.org/rfcs/rfc3330.html

Re:127.0.0.1 doesn't have an 8-bit mask

[Guppy06]

"127.0.0.1/32"

Then what's the network address? The broadcast address? Thous shalt not have a netmask longer than 30 bits.

Just like toothpaste is more like it

[Anonymous Coward]

I'm not sure we'll ever completely run out of ipv4 addresses... think of it like toothpaste in a tube. We never really run out, it just gets harder and harder to extract a large enough quantity to become useful.

Ipv6 will eventually become *easier* - and it's at THAT point that we'll all clamor aboard the train.

Re:Great, IPv6, an insecure protocol

[Anonymous Coward]

Another would be the IP management software from Lucent, VitalQIP, but it has supported IPv6 for a while now.

From your friendly neighborhood grammar nazi

[zsau]

IPv6' own automatic configuration and discovery

I'd normally write this off as a typo, but I've too often seen an apparently deliberate omission of the "s" in "-'s", which forms the English possessive. In fact, a major bank is currently having an advertising campaign in their branches involving great big banners that use "-'" where they mean "-'s". I'm therefore not aiming this post at you, jd, but at anyone and everyone who happens across it.

In English, the possessive is spelt "-'s" in almost every case --- the only expections is if the final -s is serving double duty as both the plural and the possessive, and optionally in a few primarily biblical names like Jesus or James. Thus: "James' (or James's) house's entrance's address's IPv6's" are all correct if we are referring to something owned by one instance of James, house, entrance, address or IPv6. On the other hand, the following are correct if we have multiple instances: "houses' entances' addresses'".

But it's important to note that in spite of the fact that in English individual words don't always bear a particularly strong relationship to the spelling, when you combine words or inflections, they almost always do. "An" is only used if the next word begins with a vowel in pronunciation — sometimes, there'll be an unwritten consonant before the vowel, or a consonant will be silent. "-es" is used for the plural if the word ends in a sibilant (a hissing consonant, like "(dre)ss (ja)zz (ma)tch (bri)dge"). If English isn't your native language, then you really must learn the pronunciation first, before you can become a good speller.

Is address-blocking by RIR possible?

[knorthern knight]

From some IPV6 tutorials, I understand that the RIR is identified in the header. Does this mean that a simple iptables rule could allow me to block all of AFRINIC or LACNIC or APNIC? If so, bring it on.

What about the US Gov?

[ontheroll]

While we are looking at your list, and before we complain to Ford, HP and the likes, what about the US Government? As far as I can tell from that list, a hefty chunk of addresses sit comfortably at various US Government agencies (8 of those, if I am counting right). Same goes the 2 held by the UK government.

I plead insanity

[jd]

I regard "s's" to almost always be a crime against humanity, names included, so "James'" is valid but "James's" makes me want to hurl. Most of this was very strict English lessons at school, where "s's" was punishable by death, no matter what the context, and some from rapidly figuring out that many rules were based on phonetics, not spelling. You rapidly learn shortcuts when teachers think nothing of throwing scissors or chairs. (No, Ballmer didn't go there.).

IPv6 = ewww

[4D6963]

Before I actually read what IPv6 was about years ago, I assumed it would be like the phone number extension we had in France (we added a two number prefix before all phone numbers) and that say 62.165.135.248 would become 0.0.62.165.135.248 (or a well-thought and defined prefixing scheme) and that new IPv6 addresses would be assigned to full xxx.xxx.xxx.xxx.xxx.xxx adresses, and I thought that sounded great.

Only I read about what the IPv6 would really be and I knew I didn't want it. Why? Because it's more complicated, as in much longer, and it uses hexadecimal numbers, and well, it's too unlike IPv4. Why couldn't IPv6 be a mere extension of IPv4? Why does it have to be so different? Wouldn't it have been more widely adopted if it was a more simple evolution?

Re:IPv6 = ewww

[Just Some Guy]

Only I read about what the IPv6 would really be and I knew I didn't want it. Why? Because it's more complicated, as in much longer, and it uses hexadecimal numbers, and well, it's too unlike IPv4.

That must be a major PITA for both of you who don't use DNS.

MOD PARENT FUNNY

[Anonymous Coward]

C'mon... Grandparent wrote "right" instead of "write". Answering sensibly with the word "left" is genious!

Throw parent a couple of mod points!!

[Anonymous Coward]

THANK YOU!!! Mod parent up!!

How do you turn it off if you don't care to have a user account?

ONE THING I FORGOT TO MENTION

[Anonymous Coward]

Hey Taco,

You also do realize that once you submit a post, YOU START FROM SQUARE FUCKING ONE AND HAVE TO LOAD THE WHOLE DISCUSSION AGAIN TO CONTINUE READING

Lameness filter encountered. Post aborted!
Reason: Don't use so many caps. It's like YELLING.

THAT'S BECAUSE I AM YELLING!!

Routing != Addressing - Get addresses from ISP

[billstewart]

IPv6 has at least half a dozen different address-allocation modes, though I haven't kept track of which are popular for local use and which are deprecated. If you're going to use IPv6 internally, you can go pick one. Alternatively, find an ISP that supports IPv6, either directly or through tunneling, so you can reach the rest of the IPv6 world.

More importantly, the model that the IPv6 folks want you to follow is for end-users to get address space from their ISPs, who can handle addressing hierarchically, rather than rebuilding the IPv4 Class C Swamp, where everybody not only has an address block that belonged to them, but insists that every public router in the world needs to know how to reach them and people who have large address blocks split them up into multiple parts they advertise for traffic-engineering purposes. That's led to the BGP4 address space expanding rapidly, to the point that popular large Cisco switches that can route 244000 address blocks are running out of content-addressible memory. It's not a perfect model - there's still no good solution for companies that want to have multiple ISPs for redundancy - so you may need to get your own space if you're big enough. But for what most people are doing, it's fine.

Re:Routing != Addressing - Get addresses from ISP

[ACMENEWSLLC]

Bingo. We have two links to Verizon, one to AT&T, and an ATM 100Mb/s to a local ISP which BGP's to three ISPs. We are not an ISP. We're an end user.

We do have BGP. So how do we go about getting routable IPv6 space? We can't. If I'm going to IPv6, I'm putting routable addresses on every device. That's one of the *features* of IPv6. No NAT translation needed. I'm not going to assign a private IPv6 structure now, only to re-do this later. Even if we can't currently route this out over the Internet without translation, I'm not going to design it wrong from the beginning. Once it place, this type of infrastructure is typically hell to change.

So if they want us to start implementing IPv6, they have to allow us to get address space first.

The New Shiny Tubes are Over There -

[billstewart]

That's basically what this does, is lets you point users to your new shiny IPv6 tubes instead of the old IPv4 tubes.

But yes, IPv4 has already been saved by a variety of ingenious (or evil) tech solutions in advance of running out of addresses. We've got CIDR, and proxy firewalls, and Variable-Length Subnet Masking, and NAT, and Many Ugly NAT-Traversal Solutions, and HTTP1.1's virtual hosts (which let you have multiple web server domain names at a single IP address), and SMTP's virtual hosts. We've done that already.

We're now reaching the point that if you don't switch over to IPv6 soon, you're not going to be able to get a real IPv4 address, but instead you'll be stuck behind NAT for everything.