said by whfsdude:said by leibold:There are now (and probably for a long time to come) sufficient IPv4 addresses to ensure that any business will be able to provide Internet services to the large number of IPv4-only customers in the world.
This is definitely not the case.
What I said is definitely true today and will certainly be true next year too. My guess is that it will still be true 5 years from now (can't prove that, but time will tell). Chances are by the time a business can't obtain an IPv4 address for their Internet presence anymore it won't matter because everybody has IPv6 access.
I had clearly separated businesses (providing services on the Internet) and residential users (primarily consuming those services). The quoted statement refers clearly to businesses providing Internet services while your response refers to end users. I don't disagree with what you said about ISPs not being able to assign global IPv4 addresses to each end-user but I fail to see how that relates to the quoted statement of mine ?
The allocation of IPv4 addresses has been (for the most part still is) a multi-tiered hierarchy:
- at the top you have the IANA and there is no disagreement that they are out of IPv4 addresses already (special reserves excluded, some of which may not be useful for technical reasons).
- at the next tier you have the RIRs. APNIC (Asia Pacific) was projected to run out already (hasn't yet as far as I know, but it is getting close) while AFRINIC (Africa) and LACNIC (Latin America) have reserves for about 2 years. RIPE (Europe) and ARIN (North America) were earlier projected to run out this year but may make it until next year.
- at the next tier you have the ISPs (technically this could be considered two tiers because there are a number of small ISPs that get their IPv4 allocations from large ISPs whose services they resell to consumers). Under the allocation guidelines of the RIRs ISPs were able (and in fact encouraged) to request not just the immediately needed IPv4 addresses but to make growth projections and to obtain larger blocks (for more efficient Internet routing). This means that at the time of an RIR running out of IPv4 addresses the ISPs in that region should have still sufficient IPv4 addresses for 6 to 12 month of projected growth.
- at the next tier you have both residential and commercial Internet users. It is unreasonable to expect any reserves of IPv4 addresses in the hands of residential or small business customers. However some enterprises that joined the Internet early on were allocated IP blocks that are far larger then their actual needs.
As RIRs start to run out of IPv4 addresses expect to see ISPs to start reclaiming previously allocated but now unused IPv4 addresses. I know at least one ISP that did that already years ago but for many there was no incentive to do so since a new allocation from the RIR was cheaper then tracking down the records about old allocations. Some older ISPs have used class A/B IPv4 blocks for their internal use in network operations. I wouldn't be surprised to see some of them carve chunks out of that space to allocate for customers willing to pay a premium for static IP addresses once IPv4 get scarce.
In parallel you will increasingly see IPv4 address block trades outside of the traditional IPv4 allocation hierarchy. This has already started and the first private IPv4 address sale that was made public (Nortel to Microsoft) put a price of $11.25 per IPv4 address. Since this is less then the amount some ISPs charge monthly for a static IPv4 address consider this price to go up dramatically (great investment by Microsoft!).
IP address renumbering is a big hassle but I'd be very surprised if there will not be a number of enterprises that carve up their old class A/B IPv4 blocks and sell most of it for the highest bidder. Even at only $100 per IP address selling 80% of a class B block would be over 5 million dollars and I expect the price to go well beyond that.
Once the value of an IPv4 address has risen sufficiently expect to see trades from Latin America and Africa to the Asia, Europe and North America regions.
End-users (consumers of Internet based services) may soon not be able to get global IPv4 addresses (they will still get IPv4 Internet access but it will be double NAT like AT&T is setting up right now) but any business intending to provide such services will have no problem doing so (it is likely to get more expensive but what else is new?).
Edit: APNIC is at the last /8 block which means that stricter allocation rules apply. As of today they still have 92% of this last block available for allocation.