Remember the Greek myth about Sisyphus, who was cursed to roll a huge boulder up a steep hill and then watch it roll down again for eternity? The story brings to mind IPv6, the long-anticipated upgrade to the Internet’s main communications protocol, which is known as IPv4.
Jim Bound, Chair of the North American IPv6 Task Force, CTO of the IPv6 Forum and a Senior HP Fellow, has been pushing the IPv6 boulder up the network hill for more than a decade. Bound doesn’t think the task of promoting IPv6 is as unending or pointless as Sisyphus’ boulder, but he does joke that IPv6 may not get deployed during his lifetime. (Read about “How the feds are dropping the ball on IPv6.”)
Carolyn Duffy Marsan spoke recently with Bound about the status of IPv6 adoption in the United States. Here are excerpts from the conversation:
The IPv6 community is now talking about 2010 as the date for IPv4 address exhaustion. Is that date real?
It is very real. It could be plus or minus one year. The consensus is that we have two to four years until we run out of IPv4 addresses. It’s hard to predict how the usage will work. It’s been sporadic over the years. If you go to the IPv6 Forum Web site, you’ll see a counter that shows how much time we feel is left.
What impact should the 2010 date have on corporate network managers?
The first thing they have to ask themselves is if they have enough public IPv4 addresses to sustain their businesses today, which probably they do. But do they have enough public IPv4 addresses to sustain their businesses forever? There will be new services, new users and new reasons for end-node addresses that require public IP addresses, not private IP addresses. Once the IPv4 address pool is gone, they will not be able to get a public IP address unless they move to IPv6. That’s why the registries have said it’s time to upgrade to IPv6 now.
What should U.S. companies be doing about IPv6?
It’s time for them to begin planning the move to IPv6 for their IT departments. There’s no one-size-fits-all transition plan. The first thing is to upgrade the infrastructure. You need to get your network plumbing in order so that IPv6 can co-exist and be interoperable with IPv4. Things like e-mail and Web services can be moved to IPv6 over time. You also need an address space plan. The exhaustion of the free pool of IPv4 addresses doesn’t mean the Internet melts down. The Internet will keep on going. What changes is the ability to get new public IP addresses. So anybody who is growing, whether they are a broadband provider, a manufacturing concern or a big consulting or Wall Street firm, is not going to be able grow with public IPv4 addresses.
If that’s the case, why do I know of only one U.S. company -- Bechtel -- that is rolling out IPv6?
There’s an assumption by the press that people are going to talk about their IPv6 adoptions. That’s not necessarily true. There are companies that may be on the cusp of moving to IPv6 who are not going to share that with me or you. This is very confidential stuff for some IT departments, whether you’re in banking or at Wal-Mart, because it involves security. It’s also a competitive advantage. Internet-based services are the new revenue stream. So everyone is playing IPv6 close to the vest.
U.S. federal agencies must meet a mandate to be capable of supporting IPv6 on their backbone networks by June 2008. But carriers tell me that only 10% or 15% of federal agencies are actually buying IPv6 service. Are agencies going to meet the IPv6 mandate without planning to use it?
Every agency is in the process of meeting the mandate. There’s a clear call for IPv6 to be supported in production backbones by June. I know of two agencies that are preparing for production IPv6 that I cannot reveal. You have to do the infrastructure first. You can’t get the benefits of the larger address space without the infrastructure. The June 2008 mandate says that backbones must be capable of IPv6. I think once the infrastructure is there, the need to go to IPv6 will be there.
Do you sense a lack of urgency about IPv6 in the federal market?
Even when the free pool of IPv4 addresses is depleted, the Internet keeps on ticking. Likewise, the government can keep on running their networks. What’s going to make them move to IPv6 is when they need IPv6.
How important was the federal IPv6 mandate in driving the U.S. market toward IPv6?
It was one of the key market drivers for getting all of the network hardware and software vendors to make their products IPv6 capable. There was no hope of IPv6 being deployed unless they supported it.
Federal agencies have lots of IPv4 address space, so they don’t need IPv6 addresses. What will motivate them to use IPv6 after they meet the mandate?
The Defense Department has seen the benefits of IPv6 and can articulate them beyond address space. They see many architectural benefits for mobility. Agencies like DOD that are dealing with massive, serious, life-and-death requirements see the benefit in end-to-end communications. What will motivate the rest of the federal agencies is that everyone around them will migrate to IPv6. The rest of the world doesn’t have the luxury of lots of IPv4 addresses, so they are going to IPv6. If you believe in global trade, that’s going to be a motivating factor. How will an agency deal with issues when the other party across the ocean has IPv6 and they don’t?
Are you surprised that we are six months shy of the federal IPv6 mandate and it’s not more widely deployed?
I’m totally surprised that it’s not deployed more widely. I don’t know if it was the dot-bombs or the collapse of the telcos in the early 2000s. A lot of things impacted the economy at that time that slowed down IPv6. But the situation with IPv6 is: We really need it now, and we don’t have a lot of it deployed.
Did the IPv6 community blow the federal mandate out of proportion in terms of how it would drive IPv6 deployment in the U.S.?
Yes, I think we did. The mandate said agencies had to be IPv6 capable. It didn’t define IPv6 capable in terms of running IPv6. [The U.S. government] doesn’t have a time frame for running IPv6. But I do think that in June 2008 we’ll see some federal production backbone networks running IPv6.
Do you think IPv4 address depletion is going to put more pressure on network operators to upgrade to IPv6?
Yes. There will be commercial entities that need to grow and use public IP addresses and can see the benefit of end-to-end operations to create new revenues and new growth. I believe that is in process more than we know publicly. When they need IPv6, everybody is going to want it right away. The depletion of IPv4 addresses is going to be a wakeup call.
You know, I’ve been working on IPv6 for 11 years. It’s not easy. It just isn’t a sexy thing. It’s plumbing, folks. It’s Layer three. We did such a good job of making Band-Aids for IPv4 and making it so IPv4 and IPv6 can co-exist, that everyone is waiting until the last minute to upgrade to IPv6.