Can you characterize the scaling problems of IPv4
The IP V4 scaling problems fall into two main categories. One of them is IP address space.IPv4 has a 32-bit address space. IPv6 moves to a 128-bit address space, vastly improving the number of IP addresses available. Another scaling problem with IPv4 is that Internet backbone routers have a lot of routes in them causing it to take a long time for the routing to converge. We address that in IPv6 through a more hierarchical routing system. Basically IPv6 is a new version of the Internet protocol intended to permit the global scaling of the Internet, and to make the Internet easier to use and easier to administer. What is IPv6?
IPv6 is the new version of the Internet protocol which is used to transport traffic on the Internet today. The current version is IPv4, and that version is facing a number of scaling problems that will prevent it to scale to a truly global Internet. IPv6 was begun about 10 years ago to solve those scaling problems. What can IPv6 do for network administrators?
Network administrators, are no longer going to have to administer configuration information on hosts, or set up servers to do that for them. They're going to configure the routers and then all of the end nodes of the network will be auto configured using IPv6 host auto configuration. How does IPv6 differ from IPv4?
In addition to the bigger addresses we have also tried to build an easier to administer product. One of the things that is going to help with that are the plug and play capabilities of IPv6 which include address auto-configuration. Address auto configuration means that hosts on the network, not routers, can generate their own address information using locally generated identifiers and prefix information supplied by routers. This means that there is actually no need to assign an IP address to each host any longer or even have a server like DHCP that does that, this will happen automatically. Other types of auto configuration capabilities include the ability to automatically discover servers such as a DNS server, and the ability to automatically delegate assigned routing prefixes to a home network or a small office. How far off is IPv6 from widespread adoption?
Some of the major ISPs are already providing dual-mode IPv4 IPv6 networking, but a big factor in this happening is Microsoft, and other vendors, turning IPv6 on by default. I believe it's scheduled to happen within the next one to two years, but I don't work at Microsoft so I can't make commitments on their behalf. My guess is that you will start to see some wide scale use of IPv6 in the U.S. in the next year or two although people may not be aware of it. What companies are serious about developing IPv6?
There are a huge number of companies that already have IPv6 products out. The biggest of them is Microsoft, which includes IPv6 on (Windows) XP. It isn't turned on by default, but it's there on every CD of XP that's shipped. Sun has it. Every workstation vendor has it. Equipment manufacturers such as Cisco, Nortel, Nokia, Eriksson, Siemans do. The list is very long. Also software suppliers such as Wind River are very serious about delivering these products to enable vendors to have this technology. Where is IPv6 in terms of the development standards?
The core IPv6 standards are at draft standard status in the (Internet Engineering Task Force) IETF, which is the level that indicates that the IETF believes that they're ready for widespread commercial adoption. There is one spec in the core standard that is not. It's being reviewed in the (Internet Engineering Steering Group) IESG to move to that status. We're right on the brink of having our entire set of core standards ready for widespread commercial adoption. How would you characterize the degree of difficulty in upgrading an IPv4 network to IPv6?
For a network administrator I think ultimately it will be trivial. It will be through natural replacement of their hardware over the next couple of years, or hardware they may already have, because vendors have been shipping for some time hardware that is capable of running IPv6. It will be through software upgrades, they're going to get software on those devices that run IPv6. The whole goal of this movement for the past ten years is to create a technology that is an issue of turning it on. Essentially this is flipping a switch in the routing infrastructure to turn on IPv6 routing capability which in many cases is already there.
FOR MORE INFORMATION:Best Web Links on Networking and System Management> Does the U.S. stand to lose out if other countries or other regions around the world adopt IPv6?
I think that there is a very real risk that the U.S. will not take the deployment of IPv6 seriously enough. Not from a government standpoint, but from a vendor standpoint. What we'll see is that the vendors that emerge from Asia and Europe who have IPv6 capability gain a lot of worldwide market share in the network infrastructure market. Personally I'm afraid that the North American vendors will not emphasize IPv6 enough and that they'll lose market share against Asian and European competition. Will IPv4 ever be completely replaced or become obsolete?
Not in my working lifetime. It's very hard to replace technologies. It's expected that the backbone will carry both kinds of traffic for a very long time. IPv4 will exist for a very long time.