New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
How long till we actually run out of ipv4 addresses, or there price starts getting outrageous?
dragonballz2k
Member
in General
I been wondering what happens when we get really low ipv4 addresses will vps provider be forced to do nat?
Comments
The price is probably already outrageous, and we're already low. NAT, no. We go the IPv6 route.
But I can get a BuyShared for $5/year and then get a dedicated IP
True, but soon enough prices'll start skyrocketing.
Maybe for others but for us we've already let people know that all current pricing is locked in for the life of the service.
If we hit a point where we're out of IP's we'll just stop offering that product, or go back to stock limits.
Francisco
I wouldn't bet my money on the "soon enough".
Because there's little need to adopt it. Majority of local ISPs haven't implemented IPv6 or are only starting to.
Another point is providers can still obtain IPv4 with relative ease. It may cost more, but the cost is just passed to the customer.
It's all about demand. Demand for IPv4 will never go away. It's a set resource that will never disappear.
2014 or so?
Well the design process was managed incredibly badly too. DJB's 2002-era essay still holds up for me:
http://cr.yp.to/djbdns/ipv6mess.html
Calling something a mess and citing djb in the same statement is incredibly amusing.
Well sure, if you have a 2002-era understanding of IPv6 yourself.
I'm torn on this. If there is a decent solution to allow continued access to the IPv4 network from IPv6, then IPv4 could conceivably stick around forever.
For now, IPv4 is still king, and as the shortage becomes more acute, it would be safe to assume the price will continue to rise. But the rising price will probably result in more IPs being made available as companies would then be able to make some money.
There are still 10s of not hundreds of millions of IPs that are not in use. How much is too much for 1 IP? In Asia, people already readily pay $5-$10 /IP per month (or $2 /IP for a /24). If prices get that high for RIPE or ARIN IPs, then you can be sure those with a large number of IPs will be leasing them out
In such an instance, probably all cheapo VPS with public IPv4 addresses will cease to exist. They would then all become IPv6 + NAT IPv4.
Use of IPv4 addresses would generally become more efficient, and this will probably prolong the life of IPv4.
But with the rising cost of IPv4, it would also be conceivable that a new wave of low cost IPv6 services will begin to rise, and this in itself may stimulate further uptake of IPv6. But how long will this need? Who knows?
We're far from running out. That's just alarmist news. My little local ISP has lots of extra because they don't actually need more than a few since they've been using CGN from the start.
And then there's MIT recently giving up half their /8.
Which is no insignificant thing indeed. That's around 0.5% of the entire address space, or around 16,777,216 addresses. That's a lot of IPs. Enough for many countries. I would imagine that a /9 or possibly even a /10 would be enough for the entire Baltic States.
What is the difference between 2017-era understanding vs 2002-era understanding of IPv6 @rm_? Is there anything DJB mentioned back then wrong today?
If you believed people on LET, about 6 years ago. This community has a terrible reputation for predicting the future
Also "predicting" the past... If an AI used this forum as its dataset it'd probably just answer $7 to everything then start a bidding war with its self.
I think someone will come up with an accepted method of extending IPv4's usefulness before IPv6 becomes mainstream, plenty of conceivable ways to allow the same /8's (and component parts) to be used more than once, all of which are a lot easier to implement when looking at IPv6 adoption as a comparison.
US Department of Defence has at least 218 103 808 IPs or 13 /8 blocks.
Expect offers from vps.gov
going rate is ~$10/IP - still not outrageous.
And zero motivation to change because, well, it's the government.
Private enterprise, on the other hand, will constantly do the math and free the IPs as soon as it makes financial sense.
ipv4 won't go away in your lifetime. It might be less common, but I suspect it'll live on in several forms. If nothing else, vast private networks have no motivation to renumber. 10.x is universal behind the firewall and there's no incentive for that to go away...while the world may have a problem, relatively few organizations need more than an /8 internally.
MIT is selling 8 million IPv4 to fund its IPv6 conversion. According to this article, Amazon has already purchased some of them.
http://www.networkworld.com/article/3191503/internet/mit-selling-8-million-coveted-ipv4-addresses-amazon-a-buyer.html
Because MIT doesn't have money..
That's my take on it. You have a lot of these older corps and such on old routers that have some legacy blocks and absolutely no reason to spend the money to upgrade.
I made a prediction back on one of the older LET boards that said we'd see most end user ISP's move to CGNAT with dedicated IP's being a costly addon or requiring a business plan of sorts.
Francisco
Sounds reasonable. I mean most people sit behind NAT anyway. Who honestly connects their computer's directly to the internet? Virtually no one. Everyone connects their routers to the net and then wifi to their router. Another layer of NAT won't be an issue and as long as people can still use the services they want to use, then most people won't care.
Yes it does seem like IPv4 will be around for a long while. Makes me wonder how much they will be worth in 10, 20 or 40 years? Maybe it will make sense for a bunch of hosts to stop hosting altogether, and just lease out their IPs and live off that
How would websites handle ip ban's with CGNAT it seems it would really easy to ban innocent people?
Well, CGNAT should always have IPV6 involved and most operating systems favor IPV6 over IPV4, so if the site is dual stacked it'll favor the non CGNAT way.
Francisco
It's like someone once said and I've been repeating ever since:
I'll switch to IPv6 when IPv6 can do something that I need, that IPv4 cannot.
Right now IPv6 only does these things for me:
It's a cost and benefit analysis. Benefits have to outweigh the cost, and the costs we're used to have historically held a weight significantly lower than the benefits. Just think about what you've paid for IPv4 space in the last decade vs what you've made in profit from selling it. It's cost has always been lower than it's benefits, and despite price increases it remains lower than it's benefits.
Once the price reaches a point where you can't use IPv4 to provide services that you can sell for a reasonable profit, IPv6 suddenly becomes a little more popular.
But all that assumes that an IPv4 market doesn't self regulate. How you feel about that relates mostly to your economic preferences and perspectives. You'll usually meet two kind of economic commentators who stand out:
The problem with #2 is that they get scared when markets reach unfavorable conditions and attempt to intervene before #1 can actually be proven false. Self regulation does not imply favorable conditions at all times, and sometimes it means markets collapse and have to rebuild themselves. The IPv4 market will be no different. You will have alarmists and you will have free market cheerleaders. But at the end of the day if IPv4 remains largely unregulated once owned, it's market will begin to take a more clear shape over time.
As a firm believer in market self regulation, I propose that IPv4 will fluctuate significantly in price for at least 2-3 more decades as its new market begins to take shape. It will level out in the years following the major holdouts finally cashing out on the market like MIT has done.
I'd much prefer a focus on IPv6.
@jarland
While I see what you are trying to say, there is a crucial mistake in your analysis. So, you are half-way there.
The verb regulate has 2 different meanings in these contexts. The slight of hand use of words is the source of confusion.
a) In the context of a free market, market regulation means "does this thing operate? does it fulfill its role at connecting sellers with buyers". That's it, that's all there is to it - does the market operate in accordance to its patterns that fulfill its role. "Self-regulated free markets" is redundant to simply "free markets".
b) In the context of a controlled market - the mixed system - market regulation means "does the market operate in the way the authority in charge of it wants it to? does it produce results in the way the central designer wants?"
These are 2 completely different things, and practically orthogonal. The free market regulation (i.e. how it works) does not happen in accordance to the arbitrary wishes of outside actors. In so, the free market can never "self-regulate" in the sense that the central designer of the mixed economy means the phrase "market regulation".
I agree I just think it lands on certain ears better that way. Too many people believe free market means no regulation, but you and I know that it's just a different kind of regulation, so it pushes beyond that preconceived notion and skips a conversation step by making the initial preparation to invalidate a predicted response. It's all about creative phrasing.
But yeah the two are at odds, I wouldn't say they aren't. Those who don't believe markets can level themselves out (as I call "self-regulation") jump to alarmist and intervention mode before it can be proven that markets cannot level themselves out (or "self-regulate"). The 2008 bailout is a perfect example. We intervened, we'll never know that a better reality couldn't have arrived as a result of rebuilding the market from the ground up. We just assumed it couldn't, or that we weren't willing to deal with the fallout even if it was temporary.
But the IPv4 market is one of less economical impact and it has a real chance of being left to it's own devices, I propose.