The debate whether deployment of Internet Protocol Version 6 (IPv6) should be on U.S. companies' radar screens likely won't end anytime soon.
Regardless, a pair of analysts from Burton Group, a Midvale, Utah, research firm, suggest companies start working in their labs and preparing.
Mike Disabato, Burton service director, and Daniel Golding, senior analyst, recently battled back and forth on the topic of "if or when" companies should be deploying IPv6. In the end, the duo agreed to disagree, but both said companies should start making arrangements that would ease a future deployment, such as planning a transition in labs and making sure any future hardware or software purchases are IPv6 ready.
According to Disabato, IPv6 is "ready for prime time" and companies should prepare for deployment before all IP addresses under version 4 are depleted, which by some estimates could be as soon as 2009. "It's no longer a matter of if [current IP addresses will be depleted], it's a matter of when." He added that IPv6 could provide enough IP addresses "to put one on every grain of sand on the planet."
Golding, however, battled back, citing different research that indicates current IP addresses will last at least another 20 years. He cautioned companies to resist a hasty, unplanned deployment, which would be costly.
The U.S. government and the Department of Defense, two of IPv6's strongest proponents, are estimated to spend anywhere from $25 billion to $75 billion to transition. For the average enterprise, a transition would mean buying new hardware, new software and new firewalls and rewriting applications that are supported by version 4, but would not be supported by IPv6. A transition would also put security into question, since firewalls and intrusion detection systems are not yet v6 compliant.
"I personally would not be comfortable going backward in security capabilities," Golding said.
"With proper use and proper conservation, we can go to 2026 without running out of v4 IP addresses," he said, adding that current addresses that aren't in use could be reclaimed and reused.
Many companies are also continuing to use Network Address Translation (NAT), which translates IP addresses and lets large companies use one single IP address. Currently, IPv6 does not adequately support NAT, though Golding said it eventually will.
"We're going to see IPv6 NAT whether you like it or not," he said.
But NAT is a technology that everyone loves to hate. Disabato said getting rid of NAT would be beneficial. "NAT is evil. I'd be really happy to get rid of NAT, to kill it off completely," he said.
"We've moved past the technology," he said. "It's kind of like the space shuttle. When it was introduced in the 1970s, the technology was really great. Today, not so much."
Silvia Hagen, CEO of Switzerland-based Sunny Connection AG and a SearchNetworking.com expert, agreed that companies should be prepared, but not necessarily looking at immediate deployment.
"IPv6 will come, it is inevitable," she said. "Enterprises have no need to deploy today, but they should start building expertise -- labs, education -- watch the market, put IPv6 as a requirement on shopping lists." She also cautioned companies not to "invest big money into fixing or extending IPv4, but rather put that same money into IPv6 if possible -- and doing this, they will save a lot of money in the long term."
Hagen, who is also a founding member of the Swiss IPv6 Task Force, said a possible IPv6 deployment should depend on a specific company's needs. Internet service providers and vendors, for example, should be ready now.
Hagen said companies should consider switching to IPv6 sooner than later if:
- They want to gather experience while it is not business critical yet.
- They need to extend a NAT and can do it well with IPv6.
- They run out of IPv4 address space.
- They need end-to-end security they can't get through NAT.
- The want to deploy VoIP and stumble with a NAT.
- They want to use an application that uses IPv6 features and is not available for IPv4.
- They want to use mobile IPv6.
- They need to upgrade their backbone and switching hardware anyway, and can use that as an opportunity to turn on IPv6 at the same time.