running out of IP numbers...

IP numbers are of the form q.x.y.z - 0 to 255…therefore there are potentially (256^4) or 4,294,967,296 possible machines.

What happens if we run out of numbers???

Phouchg

NO! I said POSSE!!!

Then you’ve hit the upper limit of the network nodes on the internet, and your spiffy new Java-enabled refrigerator won’t be able to notify the grocery store you’ve run out of Jolt. :wink:

That’s what IPV6 is supposed to address, by quadrupling the address information. Implementation of IPV6 was supposed to happen Any Day Now, but considering the mass of non-IPV6 compliant gear out there, and the cost of the transition, I’m not holding my breath. Classless IP (click this link for details) and reclaiming unused address ranges should get us by in the interim.


A cubicle is just a padded cell without a door.

Actually it’s a little worse than that, for a few reasons that get kindof technical. Class A addresses (16M addresses each) were given out years ago, and used extremely inefficiently. Subnetting requires wastage. When someone (typically a company) gets a block, they come in set sizes, not all of which get used, so there’s wastage there as well.

A few things have happened to keep us afloat for now.
[ul][li]CIDR makes block sizes more flexible.[/li][li]NATing (Network Address Translation) is a pretty standard technique, so that from the Internet, a site looks like a handful of addresses, but inside they have hundreds or thousands.[/li][li]The majority of Internet users are dial in ISP folks who get dynamic addresses, so that 1 address can support (typically and on average) 20 users.[/ul][/li]
People have been screaming about this for years. About 7 years ago, many people said IPv6 was necessary right away or the sky would fall. I didn’t then, and I don’t now. I suspect IPv6 is a long time away, perhaps even 10 years. The change over is complex because you need it at the clients, the services, and all the intermediate router hops. The problem is that no clients are going to change until all services and routers support it. Likewise, no service is going to support it until many clients and all routers support it. Same with Routers.

Since IPv6 sounds like IP (it’s just a new version, right?), people assume it’s a minor change. It’s really a completely new protocol with no interchangeability. It would have all the complexity of running IPX or AppleTalk over the Internet.

What billehunt said.

To remedy the situation, the powers-that-be (RIPE for us Europeans) are getting very possesive (sp?) of the addresses. Although that’s a Good Thing in principle, it’s still bloody annoying when you’re trying to get someting done in a hurry.

And of course, if addresses are allocated in very small chunks, routing tables grow explosively.

NAT technology is getting smarter & smarter - that helps.

As for IPv6: It’s going to happen, but Ghod knows when. One router company got in serious trouble by using a lot of ressources to develop the worlds first IPv6 router - only to realise that just about nobody wanted to run IPv6, at least not yet.

What the heck, it’s job security.

Norman

I can tell you that the view from the Pentagon has been that NAT is the way to go. It’s not so much a kindness for IP address conservation (as the six hundred pound gorilla, we tend to get the blokcs we need, and we have sevral Class B blocks available for our agnecy – severe overkill) as a security measure. It’s much better to have a private network numbering system inside your firewal, and a proxy server with a public interface, than exposing all your internal machines to deduction and scrutiny from ne’er-do-well hackers.

  • Rick