Because they have hardware sales and support contracts to fall back on. When the minicomputer market started to tank, for example, IBM repositioned the AS/400s (or whatever they're calling them now) as a bulletproof machine to run multiple dozen Linux instances on, and they provide support as a service.
Microsoft sells you a software package (a current version of Windows or Office) and then don't see another dime from you until it's time to buy the next version of that software - assuming you do that, of course. Yeah there are businesses that got sucked into Software Assurance but I can't imagine any smart small to medium sized business doing that anymore - why pay a fee that gets you "free upgrades" when the next upgrade cycle is after your contract expires?
Actually that's not quite true. If there's a security update for your version of IOS you can get the fixed version for the asking from them, no contract necessary. You have to specifically say "I have version 12.2(10) and bug report xxxxxx metions a IP DOS attack vector, I'm requesting 12.2(24)" or whatever. This includes taking you up to a new major or minor version if whatever you're on is deprecated, but again, only if it's a security related patch as opposed to a bugfix. You're stuck with whatever feature set you're on though. They won't take you from IP to IP Plus (or whatever they're calling it now) for free.
What you cannot do is buy a blanked-out Cisco device from Ebay (or company acquisition) and then just go download a firmware image for it. But then again, I've never really encountered any vendor that would let you do that.
ISP I worked for until 18 months ago never once bought an IOS license, we'd buy Ciscos off of Ebay and they generally had IOS images on them that needed updates, we'd E-mail Cisco TAC and get what we needed and be off to the races.
No they didn't. Baud=bitrate only in 110/300 bps era modems. 9600bps (V.32) modems were at 2400 baud but using 4 bits per symbol. Even in the post-1990 modems with trellis modulation the baudrate never cracked 3,429 but with V.34bis we were at 33.6kbps. That was the absolute maximum on an analog-only phone line. Anything past that (V.90/V.92) was one-directional PCM which you could only get away with because modern POTS lines are carried on a digital infrastructure.
CRTs are "outdated" because businesses want to sell LCDs. Flat and light is sexy. And LCDs sold like crazy back when the image quality was dramatically inferior to a CRT, and it took them years to catch up.
You make it sound as if this has already happened.
Done in one. You can even train ISC DHCP to give out different pools based on the primary IP address of the gateway for a particular VLAN. At that point all you have to worry about are keeping the pools "fed".
We're talking about the cat food you grab off the shelf when you have no cat food at home, the one your cat likes is sold out, you don't know enough to tell the difference between the ones you see in front of you and you're late for supper so you grab the one with the logo you recognize because it's a symptom of the human condition that things that are familiar are deemed safer than things that are not familiar.
No, the whole "my pet died even though I fed it name brand food" thing kind of blew the lid off of quality=expensive. My cats get the store-brand cat food, and they love it. If they're out of that, I get the next stuff up on the price scale. Either you have a cat that will mow through anything you put in front of her (mine) or a cat that's so finicky that it's a coin flip no matter what you buy. Under those conditions why would you pay $5 a bag for Meow Mix knowing your cat might hate it and prefer $2.50 Alley Cat instead?
Now when presented with two or three choices and they're all the same or similar price you're going to go with the one you've heard of, that's true.
I don't know if Cisco still does this but at one point you would get a plethora of accents not because of outsourcing but because they were bouncing your call to whatever call center was currently between 9 AM to 5 PM local time. So if you're in Chicago at 4 AM and you call the support line you're going to get someone in Australia. Basically allowed them to have only one shift of support spread out around the world rather than keeping a particular call center running 24/7.
This is not new - I have a CD of a symphony that has a passage that is rarely decoded cleanly by any player but the very best. Not the mostg expensive, but the best.
Out of curiosity, which CD is it and what CD players, in your experience, can play it correctly?
I loathed the xenon high beams when they first came out a few years ago. You know, those tiny very concentrated blue lights ? Leaves a retina trail that lasts for 20 seconds. I'm so glad that they are gone now. I've never heard if they were made illegal or if they just went out of fashion [snip]
Xenon HID headlights never went away. They're still standard equipment on higher end Japanese and some German luxury cars and optional on many others. If anything, their use by car manufacturers are expanding, not contracting.
What YOU'RE complaining about, and thankfully seem to have gone away, are those high intensity blue color lights that dropped into a standard headlight assembly. Like you mention, I don't know if they were made illegal or became "uncool" but I hated them as well. True Xenon HIDs have a completely different reflector assembly than standard halogen headlights and spread the light far more evenly and are actually less annoying that a slightly misaimed halogen headlight, and you probably see them all the time and don't even realize it. They look "blue" only compared to what turns out to be the yellow tinge of a regular headlamp, but at a glance just appear very white.
Actually the one I used to get people on was the "FAT Table". Here in Wisconsin the dominant "brand" on the ATMs is "Tyme" (As in, "Tyme Is Money") so everyone around here goes to the "Tyme Machine" to get cash. Which sounds even more retarded but at least is grammatically correct...
U.S. Cellular is fairly unique among readily available providers in WI in that they don't charge for incoming text - only outgoing. They've raised their outgoing prices along with everyone else, but they proudly advertise the fact that incoming text is free for everyone, even if they don't have a text plan. Any other national or regional carriers that do this?
Or you could take AP Pascal, like I did in H.S. (Class of 1992). Problem is, they pretty much had to have the math teachers draw straws to see who got "stuck" with the ten brainiacs who took the course, and you also had stuff like "Pascal has bad string handling because it's a math language" (which isn't technically true) and then you find out that it doesn't have a built-in function for doing anything other than square or square root. You had to code your own exponent function, which wasn't a big deal, but it made you wonder why they bothered with a built in half-assed function in the first place.
Also, we did it completely wrong IMHO. We learned the material that was on the A part of the exam for most of the year, which was basic structured programming, etc. Nothing out of the ordinary. Then, for the last MAYBE two weeks, we tried to learn what was on the B part, which was (mainly) stacks and pointers. For one thing, that was kind of bolted on to Pascal back in the day. For another, you can't really teach this "this is why you should care about this" in a mere two weeks. You can spit out a few examples in a vaccuum and hope that your students regurgitate them, but in our case that went so poorly that I think out of the ten of us in the class - all of which had 3.8 (unweighted) GPAs or higher, over half of us didn't bother taking the AB and of the other fraction of the class, only two of them got a 3 or higher on it, which was what the colleges wanted for credit. The rest of us (me included) were perfectly happy to "relearn" how pointers and stacks worked in a freshman college level C course.
Problem is, that was right when OO was kind of taking off, and everyone and their fucking sister was trying to bolt OO philosophy into EVERY CS course, even when good old-fashioned procedural thinking would have been fine. So you weren't learning C, you were learning just enough C so they could teach you C++...
Sometimes you just need a hammer and not a super-duper compressed nail gun.
I can't imagine how AIM or MSN are nearly as useful as Yahoo or ICQ - you can't leave an offline message for someone to get when they log on next. If they're not on, you can't talk to them. (It's possible both have fixed this fairly recently but it wasn't the case when I checked last, and even if it IS fixed, it was about a decade after the other two services had offlines - what took so damn long? I was sending offline messages to ICQ buddies in 1998.)
Is one of the "breaking standards" going to be something like a realistic MTU? 1500 bytes is technically too small for 100BASE-T, let alone something ten or even 100 times that fast. Let's do away with the jumbo frames bullshit once and for all. IPv4 can handle a MTU of up to 65KB, IPv6 can take that to something like 4.2GB.
Actually it turns out he had both an interest in and a talent for architecture, but by the time he figured this out he had already pissed away the opportunity to get the equivalent of a high school degree and the required background courses in the technical school. He was just a shitty painter.
What I want to know is why this wouldn't be obvious to the folks who came up with all this.
That's what I mean, why wasn't it obvious to them? How could they expect adoption of IPv6 to "just stop" if interoperability was permitted?
I don't know, you'd have to ask them.:) But actually I'd like to hear _any_ sort of reason out of them why they don't support this idea. The "not pure" argument I put forth above is pure conjecture WRT the IPv6 working group but a philosophy I've encountered time and time again in my line of work. In many situations where a company/department is using solution A and IT/management wants them to move to solution B, oftentimes an easy migration path between A and B is convieniently left out or made difficult so that the end users don't take the easy way out and use just enough of B to feel like they "switched" but in reality are only 80% cut over before they move on to the next project. I don't support this at all, I'm just saying I've seen it coming from others, people who in general have a problem seeing the forest for the trees.
Exactly. The nationalism and fear that put Hitler in power could have just as easily put Hans Gruber (how's that for a generic German name?) in power. If anything, Nationalism and anti-Semitism could have actually festered for longer, attracted someone just as mad with power, but who wasn't going to have an eventual descent into Parkinson's and madness. Someone who would have implemented "The Final Solution" earlier on in his reign and more efficiently, and who would have been a better general and used his military more effectively. The Germans beat themselves in some cases just as effectively as the Allies did. (Not to say the Allies didn't do anything - far from it - but Germany made some critical missteps early on that were obvious to everyone but Hitler.)
Right, I was agreeing, but it was 2 AM and I wasn't being clear.:) My point was that the IPv6 camp seems to think we'd all do the intermediate step and then "just stop" and never really adopt IPv6 completely (i.e., the internet would be hybrid until the end of time.) Because I was wandering a little in my writing I ended up repeating you/myself. Sorry about that.
Politics. Or religion, pick one. These guys assume that because in their little fiefdom, they were able to completely switch over to IPv6, that the rest of the world can too. They don't like anything that is a "kludge" that would keep the world operating in kind of a dual-state mode for some time. In their view, when it's time to switch to IPv6, it's time to switch, period. Anything else isn't "pure" enough. They assume that if they give people an "out" to avoid switching over to pure IPv6 that everyone will take it and just kind of sit there and the problem (resource exhaustion) will simply have moved to a different domain (there's still going to be the same amount of IPv4 addresses buried in the RFC2893 space) and won't be solved. This is wrong for two reasons. A) the resource exhaustion, as I mention in another post somewhere in this thread, has been "coming" for about a decade now, and we somehow always find a way to stave it off, be it switching web servers from single-IP-per-domain to host header-based hosting (what you all know as many web sites on one IP) or telling companies with 200 PCs "here's your/28, please NAT everything that doesn't need to be touched from the world" or some other vaguly kludgy but overall workable scheme. Yes, eventually this will run out, but it won't be in two years. B) if you DID use the 2893 extensions to "park" all of the IPv4 space in IPv6, you could pretty much, more or less, that day, stop handing out IPv4 space and make all new allocations move to IPv6, but it wouldn't be a problem because everyone would be able to "talk" to everyone else. If for legacy reasons you still need to give out a few more IPv4 blocks, (companies that have large allocations needing a little more space before they completely switch in a couple years, for example) you'd have the space to do that in, but once you have more than 50% of the net using IPv6 you've hit the tipping point and there's going to be less need for people to even have IPv4 legacy blocks. Eventually you just stop handing them out altogether, full stop. At all times during this sort of move, all clients could reach all servers and life would be grand. But that wouldn't be "pure".
The IPv4 crunch has been 2 years away for at least 10 years.
What he said. I've been in the ISP biz since 1995, and I've been hearing "the sky is falling on IPv4" since about 1998. I _know_ how hard it is to get newly minted IP space from ARIN or an appropriate body. They had to get tough on it when CIDR took off not just to curb IP usage, but to curb the routing table size. (It was growing like crazy for a while because everyone and his brother was allocating a/24 out of what used to be a/8 or a/16, and routers were starting to fall over.) The nice side effect of this is you can't have Joe SOHO apply to ARIN for a block of IPs - you have to be able to PROVE you can use a/20 right out of the gate (as in, you're already using the equivalent of a/20 from your upstream provider - that's _16_ class C's, boys and girls). Otherwise, you're getting your new IP space from your upstream until you reach that size. I don't have the link handy but there are still something like 65 million IP addresses available _before_ you start making places like M.I.T. or Ford give back some of their old pre-ARIN/8s. Now yes, it's true that eventually this space will run out, because it is a finite resource. But they've been getting it wrong on predicting this resource exhaustion for pretty much the Internet's entire adult life (if you figure it hit adulthood when non-academic interests were allowed access around about 1994.) And this takes into account the.com bubble, where pets were getting their own IP spaces. The instant I realized that IPv6 was a _replacement_ for IPv4 rather than an augmentation to it, I knew it was going to be a brutal, nasty switch. Hopefully we can keep the alarmists in check and work out a way to move onto the next big thing before resources become scarce - but that probably means taking IPv6 out behind the barn and putting a bullet in its brain and working out something that extends IPv4 so that the instant a new customer gets on this new network they can access everything that we know as "the internet" without having to worry about 6to4 tunnels or any of this other crap. The guys who developed IPv6 seem to think that since it was easy to change their ten PCs in the lab to the new addressing scheme that it's going to be that easy for everyone. Right now, we seem to be in a big game of chicken. No one's going to move until everyone moves, and that's a self-fulfilling doomed prophecy.
Slashdot Mirror
Because they have hardware sales and support contracts to fall back on. When the minicomputer market started to tank, for example, IBM repositioned the AS/400s (or whatever they're calling them now) as a bulletproof machine to run multiple dozen Linux instances on, and they provide support as a service.
Microsoft sells you a software package (a current version of Windows or Office) and then don't see another dime from you until it's time to buy the next version of that software - assuming you do that, of course. Yeah there are businesses that got sucked into Software Assurance but I can't imagine any smart small to medium sized business doing that anymore - why pay a fee that gets you "free upgrades" when the next upgrade cycle is after your contract expires?
That honor, in my humble opinion would be the Post's December 13, 2007 headline:
IKE 'BEATS' TINA TO DEATH
Reporting the death of Ike Turner of a cocaine overdose.
That one made me actually laugh out loud.
Sweet merciful Jesus. If I could give you 100 mod points, I would.
Actually that's not quite true. If there's a security update for your version of IOS you can get the fixed version for the asking from them, no contract necessary. You have to specifically say "I have version 12.2(10) and bug report xxxxxx metions a IP DOS attack vector, I'm requesting 12.2(24)" or whatever. This includes taking you up to a new major or minor version if whatever you're on is deprecated, but again, only if it's a security related patch as opposed to a bugfix. You're stuck with whatever feature set you're on though. They won't take you from IP to IP Plus (or whatever they're calling it now) for free.
What you cannot do is buy a blanked-out Cisco device from Ebay (or company acquisition) and then just go download a firmware image for it. But then again, I've never really encountered any vendor that would let you do that.
ISP I worked for until 18 months ago never once bought an IOS license, we'd buy Ciscos off of Ebay and they generally had IOS images on them that needed updates, we'd E-mail Cisco TAC and get what we needed and be off to the races.
Welcome to my friends list, fellow "The Tick" reader/watcher. (Which is it, out of curiosity?)
No they didn't. Baud=bitrate only in 110/300 bps era modems. 9600bps (V.32) modems were at 2400 baud but using 4 bits per symbol. Even in the post-1990 modems with trellis modulation the baudrate never cracked 3,429 but with V.34bis we were at 33.6kbps. That was the absolute maximum on an analog-only phone line. Anything past that (V.90/V.92) was one-directional PCM which you could only get away with because modern POTS lines are carried on a digital infrastructure.
CRTs are "outdated" because businesses want to sell LCDs. Flat and light is sexy. And LCDs sold like crazy back when the image quality was dramatically inferior to a CRT, and it took them years to catch up.
You make it sound as if this has already happened.
Done in one. You can even train ISC DHCP to give out different pools based on the primary IP address of the gateway for a particular VLAN. At that point all you have to worry about are keeping the pools "fed".
We're talking about the cat food you grab off the shelf when you have no cat food at home, the one your cat likes is sold out, you don't know enough to tell the difference between the ones you see in front of you and you're late for supper so you grab the one with the logo you recognize because it's a symptom of the human condition that things that are familiar are deemed safer than things that are not familiar.
No, the whole "my pet died even though I fed it name brand food" thing kind of blew the lid off of quality=expensive. My cats get the store-brand cat food, and they love it. If they're out of that, I get the next stuff up on the price scale. Either you have a cat that will mow through anything you put in front of her (mine) or a cat that's so finicky that it's a coin flip no matter what you buy. Under those conditions why would you pay $5 a bag for Meow Mix knowing your cat might hate it and prefer $2.50 Alley Cat instead?
Now when presented with two or three choices and they're all the same or similar price you're going to go with the one you've heard of, that's true.
I don't know if Cisco still does this but at one point you would get a plethora of accents not because of outsourcing but because they were bouncing your call to whatever call center was currently between 9 AM to 5 PM local time. So if you're in Chicago at 4 AM and you call the support line you're going to get someone in Australia. Basically allowed them to have only one shift of support spread out around the world rather than keeping a particular call center running 24/7.
This is not new - I have a CD of a symphony that has a passage that is rarely decoded cleanly by any player but the very best. Not the mostg expensive, but the best.
Out of curiosity, which CD is it and what CD players, in your experience, can play it correctly?
I loathed the xenon high beams when they first came out a few years ago. You know, those tiny very concentrated blue lights ? Leaves a retina trail that lasts for 20 seconds. I'm so glad that they are gone now. I've never heard if they were made illegal or if they just went out of fashion [snip]
Xenon HID headlights never went away. They're still standard equipment on higher end Japanese and some German luxury cars and optional on many others. If anything, their use by car manufacturers are expanding, not contracting.
What YOU'RE complaining about, and thankfully seem to have gone away, are those high intensity blue color lights that dropped into a standard headlight assembly. Like you mention, I don't know if they were made illegal or became "uncool" but I hated them as well. True Xenon HIDs have a completely different reflector assembly than standard halogen headlights and spread the light far more evenly and are actually less annoying that a slightly misaimed halogen headlight, and you probably see them all the time and don't even realize it. They look "blue" only compared to what turns out to be the yellow tinge of a regular headlamp, but at a glance just appear very white.
Actually the one I used to get people on was the "FAT Table". Here in Wisconsin the dominant "brand" on the ATMs is "Tyme" (As in, "Tyme Is Money") so everyone around here goes to the "Tyme Machine" to get cash. Which sounds even more retarded but at least is grammatically correct...
Wow, I got Grammar Nazied.
Since you're looking up things, try "colloquial".
Don't forget to castigate people for going to the "ATM machine" while you're at it.
(The above written mostly in jest.)
(Mostly.)
U.S. Cellular is fairly unique among readily available providers in WI in that they don't charge for incoming text - only outgoing. They've raised their outgoing prices along with everyone else, but they proudly advertise the fact that incoming text is free for everyone, even if they don't have a text plan. Any other national or regional carriers that do this?
I'm pretty sure all he would be able to say at this point is "Braaaaaaaaaaains!"
Or you could take AP Pascal, like I did in H.S. (Class of 1992). Problem is, they pretty much had to have the math teachers draw straws to see who got "stuck" with the ten brainiacs who took the course, and you also had stuff like "Pascal has bad string handling because it's a math language" (which isn't technically true) and then you find out that it doesn't have a built-in function for doing anything other than square or square root. You had to code your own exponent function, which wasn't a big deal, but it made you wonder why they bothered with a built in half-assed function in the first place.
Also, we did it completely wrong IMHO. We learned the material that was on the A part of the exam for most of the year, which was basic structured programming, etc. Nothing out of the ordinary. Then, for the last MAYBE two weeks, we tried to learn what was on the B part, which was (mainly) stacks and pointers. For one thing, that was kind of bolted on to Pascal back in the day. For another, you can't really teach this "this is why you should care about this" in a mere two weeks. You can spit out a few examples in a vaccuum and hope that your students regurgitate them, but in our case that went so poorly that I think out of the ten of us in the class - all of which had 3.8 (unweighted) GPAs or higher, over half of us didn't bother taking the AB and of the other fraction of the class, only two of them got a 3 or higher on it, which was what the colleges wanted for credit. The rest of us (me included) were perfectly happy to "relearn" how pointers and stacks worked in a freshman college level C course.
Problem is, that was right when OO was kind of taking off, and everyone and their fucking sister was trying to bolt OO philosophy into EVERY CS course, even when good old-fashioned procedural thinking would have been fine. So you weren't learning C, you were learning just enough C so they could teach you C++...
Sometimes you just need a hammer and not a super-duper compressed nail gun.
I can't imagine how AIM or MSN are nearly as useful as Yahoo or ICQ - you can't leave an offline message for someone to get when they log on next. If they're not on, you can't talk to them. (It's possible both have fixed this fairly recently but it wasn't the case when I checked last, and even if it IS fixed, it was about a decade after the other two services had offlines - what took so damn long? I was sending offline messages to ICQ buddies in 1998.)
Is one of the "breaking standards" going to be something like a realistic MTU? 1500 bytes is technically too small for 100BASE-T, let alone something ten or even 100 times that fast. Let's do away with the jumbo frames bullshit once and for all. IPv4 can handle a MTU of up to 65KB, IPv6 can take that to something like 4.2GB.
Quick reading: http://staff.psc.edu/mathis/MTU/index.html
Actually it turns out he had both an interest in and a talent for architecture, but by the time he figured this out he had already pissed away the opportunity to get the equivalent of a high school degree and the required background courses in the technical school. He was just a shitty painter.
What I want to know is why this wouldn't be obvious to the folks who came up with all this.
:) But actually I'd like to hear _any_ sort of reason out of them why they don't support this idea. The "not pure" argument I put forth above is pure conjecture WRT the IPv6 working group but a philosophy I've encountered time and time again in my line of work. In many situations where a company/department is using solution A and IT/management wants them to move to solution B, oftentimes an easy migration path between A and B is convieniently left out or made difficult so that the end users don't take the easy way out and use just enough of B to feel like they "switched" but in reality are only 80% cut over before they move on to the next project. I don't support this at all, I'm just saying I've seen it coming from others, people who in general have a problem seeing the forest for the trees.
That's what I mean, why wasn't it obvious to them? How could they expect adoption of IPv6 to "just stop" if interoperability was permitted?
I don't know, you'd have to ask them.
Exactly. The nationalism and fear that put Hitler in power could have just as easily put Hans Gruber (how's that for a generic German name?) in power. If anything, Nationalism and anti-Semitism could have actually festered for longer, attracted someone just as mad with power, but who wasn't going to have an eventual descent into Parkinson's and madness. Someone who would have implemented "The Final Solution" earlier on in his reign and more efficiently, and who would have been a better general and used his military more effectively. The Germans beat themselves in some cases just as effectively as the Allies did. (Not to say the Allies didn't do anything - far from it - but Germany made some critical missteps early on that were obvious to everyone but Hitler.)
Right, I was agreeing, but it was 2 AM and I wasn't being clear. :) My point was that the IPv6 camp seems to think we'd all do the intermediate step and then "just stop" and never really adopt IPv6 completely (i.e., the internet would be hybrid until the end of time.) Because I was wandering a little in my writing I ended up repeating you/myself. Sorry about that.
Politics. Or religion, pick one. These guys assume that because in their little fiefdom, they were able to completely switch over to IPv6, that the rest of the world can too. They don't like anything that is a "kludge" that would keep the world operating in kind of a dual-state mode for some time. In their view, when it's time to switch to IPv6, it's time to switch, period. Anything else isn't "pure" enough. They assume that if they give people an "out" to avoid switching over to pure IPv6 that everyone will take it and just kind of sit there and the problem (resource exhaustion) will simply have moved to a different domain (there's still going to be the same amount of IPv4 addresses buried in the RFC2893 space) and won't be solved. This is wrong for two reasons. A) the resource exhaustion, as I mention in another post somewhere in this thread, has been "coming" for about a decade now, and we somehow always find a way to stave it off, be it switching web servers from single-IP-per-domain to host header-based hosting (what you all know as many web sites on one IP) or telling companies with 200 PCs "here's your /28, please NAT everything that doesn't need to be touched from the world" or some other vaguly kludgy but overall workable scheme. Yes, eventually this will run out, but it won't be in two years. B) if you DID use the 2893 extensions to "park" all of the IPv4 space in IPv6, you could pretty much, more or less, that day, stop handing out IPv4 space and make all new allocations move to IPv6, but it wouldn't be a problem because everyone would be able to "talk" to everyone else. If for legacy reasons you still need to give out a few more IPv4 blocks, (companies that have large allocations needing a little more space before they completely switch in a couple years, for example) you'd have the space to do that in, but once you have more than 50% of the net using IPv6 you've hit the tipping point and there's going to be less need for people to even have IPv4 legacy blocks. Eventually you just stop handing them out altogether, full stop. At all times during this sort of move, all clients could reach all servers and life would be grand.
But that wouldn't be "pure".
The IPv4 crunch has been 2 years away for at least 10 years.
/24 out of what used to be a /8 or a /16, and routers were starting to fall over.) The nice side effect of this is you can't have Joe SOHO apply to ARIN for a block of IPs - you have to be able to PROVE you can use a /20 right out of the gate (as in, you're already using the equivalent of a /20 from your upstream provider - that's _16_ class C's, boys and girls). Otherwise, you're getting your new IP space from your upstream until you reach that size. I don't have the link handy but there are still something like 65 million IP addresses available _before_ you start making places like M.I.T. or Ford give back some of their old pre-ARIN /8s. Now yes, it's true that eventually this space will run out, because it is a finite resource. But they've been getting it wrong on predicting this resource exhaustion for pretty much the Internet's entire adult life (if you figure it hit adulthood when non-academic interests were allowed access around about 1994.) And this takes into account the .com bubble, where pets were getting their own IP spaces. The instant I realized that IPv6 was a _replacement_ for IPv4 rather than an augmentation to it, I knew it was going to be a brutal, nasty switch. Hopefully we can keep the alarmists in check and work out a way to move onto the next big thing before resources become scarce - but that probably means taking IPv6 out behind the barn and putting a bullet in its brain and working out something that extends IPv4 so that the instant a new customer gets on this new network they can access everything that we know as "the internet" without having to worry about 6to4 tunnels or any of this other crap. The guys who developed IPv6 seem to think that since it was easy to change their ten PCs in the lab to the new addressing scheme that it's going to be that easy for everyone. Right now, we seem to be in a big game of chicken. No one's going to move until everyone moves, and that's a self-fulfilling doomed prophecy.
What he said. I've been in the ISP biz since 1995, and I've been hearing "the sky is falling on IPv4" since about 1998. I _know_ how hard it is to get newly minted IP space from ARIN or an appropriate body. They had to get tough on it when CIDR took off not just to curb IP usage, but to curb the routing table size. (It was growing like crazy for a while because everyone and his brother was allocating a