One fact has become clear about IPv6, the next-generation Internet protocol developed to gradually replace the current IPv4: Adoption by U.S. enterprises is not happening on Internet time. Even those who see potential in the technology, like Dan Demeter, CIO of talent management company Korn\/Ferry International, are taking it slow. He plans to introduce IPv6 by 2010 as part of a worldwide network upgrade for his company.\n\n\u201cWe believe that [by] adopting IPv6 and restructuring our network routers and servers, we can deliver faster and more reliable and secure client solutions,\u201d Demeter says. Also, Korn\/Ferry employees use BlackBerry mobile devices to access key company executive search data, and Demeter wants to explore the potential of IPv6 for providing additional mobile services.\n\nAmong top benefits, IPv6 promises a significant increase in the number of addresses available for networked devices such as mobile phones, and simpler administration of networks.China\u2019s IPv6 Plan\nWhat\u2019s the vision for IPv6 on the other side of the globe? See Internet Strategy: China\u2019s Next Generation Internet.\n\nBut Demeter says Korn\/Ferry is in the exploration stage, with no firm time frame for a pilot test. \u201cOur approach is to focus on the areas where we can derive the most benefits and move ahead in gradual fashion as our experience grows and as we ensure that all the infrastructure components are compatible with IPv6.\u201d\n\nHe\u2019s not alone. Federal government agencies are mandated by the Office of Management and Budget to move their network backbones to IPv6 by June 2008\u2014and so are the contractors that do business with agencies. But outside that space, few organizations seem to be deploying the standard. Research firm Gartner estimates enterprise adoption at less than 1 percent.\n\nShould IPv6 be on your drawing board yet? Consider the key issues and the experiences of early adopters carefully.\n\nFew Business Drivers for IPv6\nSeveral factors are fueling the sluggish adoption rate. A study by Cisco in 2006 cited the lack of dedicated funding and IT staff for IPv6 implementations.\n\nAnother hurdle: \u201cThe fact that IPv6 implementation is viewed more as a technology issue than a business benefits driver probably also is an obstacle to its immediate widespread adoption in the U.S.,\u201d says Michael A. Gold, a senior partner in the litigation group of Los Angeles law firm Jeffer, Mangels, Butler & Marmaro and cochair of the firm\u2019s Discovery Technology Group.\n\n\u201cThis is very shortsighted in terms of global competition,\u201d Gold says. \u201cIn the not-too-distant future, many home appliances\u2014even dog collars\u2014will be Internet connected. Many automobiles are connected today. Each of these devices will require using an Internet address in order to communicate across the network.\u201d\n\nQuite simply, the system will run out of addresses some years from now without IPv6. (ARIN, a regional Internet registry organization that provides services related to the management of Internet number resources, won\u2019t comment on when those IPv4 addresses will run out.)\n\nOther countries, notably China, have pushed the implementation of IPv6 more aggressively than the United States.\n\nAmong the other possible benefits of IPv6, the technology enables a more simplified network architecture that removes network address translation devices, clearing the way for powerful peer-to-peer capabilities, says Erica Johnson, senior manager of software and applications and IPv6 consortium manager at the University of New Hampshire\u2019s InterOperability Laboratory. The lab oversees the Moonv6 project, a global effort to test IPv6 equipment from different vendors.\n\nIPv6 also includes a greater amount of usable address space for additional nodes on the network, allowing better utilization of multiuser technologies such as VoIP, interactive video and collaborative applications, she notes.\nBut Johnson concedes that even with the potential gains from IPv6, building a business case for adoption will be a challenge for many. \u201cA lot of that has to do with testing and education,\u201d she says. \u201cIt\u2019s not going to be a light switch; we don\u2019t have a Y2K effect with deploying IPv6.\u201d\n\nSome analysts are more blunt. \u201cCommercial enterprises have little reason to adopt IPv6,\u201d says David Willis, research VP at Gartner. \u201cMigration costs are very high for established IP networks, and attempts to transition even moderate-size networks have revealed many unexpected problems and hidden costs.\u201d\n\nWillis says most of the benefits of IPv6 \u201ccan be delivered with current IP [IPv4] workarounds such as network address translation and IPsec [the Internet security protocol].\u201d\n\nWillis adds that he expects IPv6 to \u201ccreep into the enterprise as we see stronger Vista rollouts in 2008.\u201d Enterprises will use various approaches to support both IPv4 and IPv6 for several years, he says.\n\nEarly Adopter Lessons\nCIOs starting to explore the IPv6 issue can learn from the approach of early adopters like engineering and construction giant Bechtel. By 2003, the U.S. Department of Defense, a big Bechtel customer, had called for departmentwide deployment of IPv6 by 2008. Bechtel began seeing RFPs from the U.S. Army and other customers explicitly calling for IPv6 products and services. So in 2004, Bechtel launched a phased, enterprisewide deployment of IPv6 \u201cdesigned to develop broad awareness and competence in the new protocol, with the initial deployment focused on our government business unit,\u201d says Fred Wettling, Bechtel fellow and technology strategy manager.\n\nThe company sees an opportunity to create an IT infrastructure that will be \u201ca platform for future innovation,\u201d he says. \u201cThis is a technology that can transform the way we do business.\u201d\n\nWettling says Bechtel sees IPv6 as an enabling technology, as the Web was in the 1990s. For example, the company is exploring how IPv6 will help with wireless sensor networks to help track logistics, and with mobile ad hoc networks that can be set up quickly at the start of a project.\n\nBechtel\u2019s IT group tried to minimize the problems and costs associated with a broad technology change by using a planned, gradual approach spanning several years. This included sending 3 dozen people to an \u201cIPv6 boot camp\u201d run by Native6 (now part of Command Information, a provider of IPv6 training and services) and creating an IPv6 lab to perform distributed configurations and testing without putting Bechtel\u2019s production network at risk.\n\n\u201cWe set up small IPv6 labs at four locations, each with a few servers, routers, switches, and put them in isolated networks within each office and interconnected them across the Internet,\u201d Wettling says.\n\nBy the end of 2006, Bechtel had enabled IPv6 on the production networks and hundreds of computers at four of its primary sites, and created a scalable model for future deployments.\n\nThe company instructed all its application developers on how to configure machines for IPv6. Today, Bechtel has more than 9,000 computers (desktops, portables and servers) in 70 cities worldwide running IPv6. The majority of its offices support IPv6, and the company is turning on other offices one at a time.\n\nHardware Hiccups\nWhat challenges did Bechtel encounter on its road to IPv6? While most of the applications weren\u2019t affected by the change in IP version, several presented problems. First, some databases weren\u2019t set up with big enough fields to accommodate IPv6 addresses and had to be expanded.\n\nAlso, not all commercial or internally developed applications have the needed IPv6 attributes in them. Some of Bechtel\u2019s monitoring and configuration software had to be tweaked to display IPv6 data.\n\n\u201cNot all products out there [such as Windows XP] have the IPv6 features we want,\u201d Wettling adds. \u201cXP doesn\u2019t fully support IPv6 as well as [Microsoft\u2019s] Vista does.\u201d Bechtel will start deploying Vista later this year, he says.\n\nFor these reasons and others, aeronautics manufacturer Lockheed Martin figures its move to IPv6 will be a huge undertaking. \u201cThe transition to IPv6 will require a greater effort than the Y2K bug,\u201d says Frank Cuccias, director of Lockheed\u2019s IPv6 Center of Excellence. \u201cRemember that Y2K only affected a subset of systems; IPv6 will affect almost all current systems.\u201d\n\nLockheed Martin, given its many government customers, began looking at IPv6 seven years ago in its labs. The company is in the midst of a pilot program to convert part of its Global Vision Network to IPv6. So far the program is progressing well, Cuccias says.\n\n\u201cWe realize that if our customers are going transition to IPv6, we need to be out in front of the technology,\u201d Cuccias says. The company launched the pilot \u201cto illustrate to our customers that it\u2019s not as simple as buying new IPv6 hardware and turning it on,\u201d he says.\n\nGovernment agencies are among the earliest adopters of IPv6, largely because of the mandate. The U.S. Government Accountability Office (GAO) completed an inventory of its hardware and software, and assessed which were not IPv6-compliant, says CIO Joseph Kraus.\n\nThe agency also conducted a network assessment of how IPv6 traffic is transmitted from the Internet and GAO\u2019s private network. \u201cAs part of a planned upgrade of our network infrastructure [scheduled for 2008], we included IPv6 compatibility in our specifications,\u201d Kraus says.\n\nDuring its prep work, the agency learned that its network service providers weren\u2019t able to transmit IPv6 packets and needed to upgrade their infrastructure. \n\nAnother potential IPv6 challenge is developing network engineering expertise, says Korn\/Ferry\u2019s Demeter. \u201cWhile IPv6 presents several advantages over IPv4, it requires the engineering and systems operations talent to design, build, and maintain the network to maximize its potential and to justify the investment,\u201d Demeter says. \u201cWe have the talent in-house, but we need the time to build the expertise in this new area.\u201d\n\nWhat\u2019s the bottom line for CIOs outside the government sphere right now? Despite the gradual depletion of IPv4 address space, Gartner\u2019s Willis sees no urgency to adopt IPv6. \u201cThere is no real driver besides the IP address shortage,\u201d he says. \u201cWhat this means is that we\u2019ll be living in a mixed IPv4\/IPv6 environment until well past 2013. Fortunately, coexistence of both protocols is easy, although it will drive support costs up while we are in this mixed environment.\u201d\n\nBob Violino is a New York-based freelance writer. E-mail feedback to firstname.lastname@example.org.