Technical
and Economic Assessment of IPV6 TASK
FORCE
General: The Task Force concludes that the United States and other economies are in the early stages of IPv6 adoption and deployment. As such, many uncertainties exist with respect to the benefits and costs of prospective market applications and, therefore, the benefits and costs of alternative transition scenarios. Nevertheless, a consensus exists with respect to the likely long-term importance of IPv6 adoption. At this time, most of the stakeholders participating in the Task Force’s activities believe that the current market-driven adoption of IPv6 by the private sector is proceeding at a reasonable pace and that the instituted transition mechanisms will enable efficient migration at acceptable cost. The push for adoption in other countries, however, could potentially change this situation, and the complexity of the infrastructure necessary to effect the transition from IPv4 to IPv6 may require additional support. Thus, a number of technology and economic policy issues need to be examined regularly in order to determine, over time, what support, if any, may be needed for the growing IPv6 activities by industry and by federal, state, and local governments. In this context, the Task Force has reached the following findings: (1) IPv6 is a complex standard consisting of a suite of protocols, definitions, transition mechanisms, and operational procedures. These protocols are at varying stages of maturity, with varying scopes of applicability and varying subsets of mandatory/recommended/discretionary implementation options. In the near term, the net benefits of IPv6 compared with IPv4 will vary among organizations and deployment scenarios, both domestically and internationally. Still, the Task Force recognizes the long-term benefits of an evolution to a protocol with a significantly larger address space than IPv4. (2) A collection of techniques (e.g., NATs) have been developed and deployed in recent years to accommodate the growing demand for IPv4 addresses, but these “fixes” impose operational inefficiencies and costs. If, as a many observers anticipate, large-scale demand for new, address-intensive applications such as mobile communications, remote monitoring, and consumer Internet-TV emerge, the continued viability of such techniques comes into question. (3) IPv6 stakeholders can be organized into four major groups: • Infrastructure (Hardware and Software) Vendors • Application Vendors • Internet Service Providers (ISPs) • Internet Users The potential benefits, costs, and risks associated with IPv6 adoption can vary significantly across this range of stakeholders. Any analysis of such issues should, therefore, be specific to each of these groups. (4) The Task Force recognizes that the complexity involved in IPv6 adoption and use also varies greatly with specific deployment scenarios; ranging from so-called “isolated greenfield” implementations (primarily new private networks) to transitioning the Internet (existing and public networks), which is a large and complex problem. In addition, the scope of IPv6 adoption can vary greatly within a network, so careful consideration must be given to which network devices, applications, management, and control functions are to be affected. (5) Based on the above findings, the Task Force concludes that both public- and private-sector users of networked information technology should begin planning for the emergence of IPv6 technologies and analyze requirements and appropriate schedules for adoption. In the near term, in order to ensure the security and stability of both new IPv6-enabled IT systems and the existing systems with which they must interoperate, the Task Force stresses that careful planning, development, and evaluation should be undertaken for the forthcoming dual-standard environment. Within federal networks, the identified need to expedite IPv6 planning and analysis is consistent with other recent government studies [ 256 ] and evolving policies. [ 257 ] (6) Recognizing that several crucial aspects of IPv6 remain to be specified and that all elements of the technical basis for the standard need significant additional test and evaluation experience, the Task Force notes that the federal government will need to commit new resources and to work collaboratively with other public and private sector entities to address these outstanding research, development, and testing issues. Given the scope and importance of these issues, identification of a specific entity to coordinate these activities within the federal government and among similar international efforts should be considered. Economic Growth and Competitive Impacts: The global scope of the Internet means that both domestic private investment and the standards infrastructure supporting the Internet must evolve in a timely manner. That evolution will be complex because of the multiple industries involved in the delivery of Internet infrastructure and the myriad of existing and emerging services that depend upon it. Although the Task Force concurs with the general view that the transition to IPv6 is occurring at an acceptable pace, some industries in the Internet supply chain are migrating to the new protocol at faster rates than are others. Therefore, a number of trends and potential barriers must be continually monitored and assessed. Major portions of the Internet infrastructure hardware and software markets appear to be IPv6 “capable” already, and over the next four or five years, the vast majority of network hardware, operating systems, and network-enabled software packages (e.g., databases, email) will be sold with IPv6 capabilities. This capability is not actually “turned on,” however. In the next few years, users will begin to “enable” or “turn on” this capability in operating systems, or they will purchase operating systems with IPv6 “on by default.” In fact, the majority of Linux-based operating systems are IPv6 enabled today, and the next version of Windows, due out in 2007, will likely be IPv6 enabled by default. As operating systems become enabled and early adopters provide “lessons learned,” respondents predict that users will start to enable routers, followed finally by applications. Applications are the key driver because they will create demand for the aforementioned categories of IPv6 infrastructure. Application vendors are moving toward IPv6 at a much slower pace than are infrastructure vendors, however. Many application developers have been testing IPv6 and planning to integrate IPv6 into their products, although very few have actually begun selling IPv6-capable products, at least in the United States. Many of these vendors are indicating that they plan to release IPv6-capable products as early as 2007. Many ISPs that do not also provide Internet backbone facilities are not offering IPv6 connectivity because they do not want to incur costs without a reasonably certain return on investment. Consequently, although numerous ISPs are currently engaged in testing activities and may offer limited IPv6 services, they appear to be waiting for a significant number of mainstream customers to request IPv6 connectivity. At that point, those ISPs indicate that they will be prepared to provide service in six months to one year. Like users, however, ISPs do not intend to offer IPv6 service until major hardware and software network components are in place. More generally, emerging and future address-intensive, peer-to-peer Internet applications will exhibit an iterative relationship with the supporting infrastructure. That is, the availability of a higher capacity and more efficient standards infrastructure leverages private-sector innovation, which, in turn, increases the use of and demand for improvements in the supporting infrastructure. Thus, monitoring the evolution of the “chicken-or-egg” relationship between infrastructure and innovation is important for long-term domestic economic growth policy and implies an optimum balance between public and private investment in technology-based industries. Most stakeholders in the U.S. believe that IETF-fashioned transition strategies will allow fast market response at reasonable cost to the emergence of demand for IPv6-dependent applications. Over time, the net benefits of IPv6 will increase for all industries using the Internet, but currently, nations competing with the United States have a greater incentive to migrate to IPv6 due to perceived limitations in IPv4 address space. That fact is a two-edged sword. On the one hand, the vast installed base of IPv4 infrastructure and applications buys time with respect to transition decisions. On the other hand, a large installed base can act as a barrier to change from IPv4 to IPv6 because of the sunk costs and the fact that the IPv4 infrastructure provides an acceptable level of service and functionality for most users. Because the rest of the world is clearly migrating toward IPv6, a long-term competitiveness issue faces the U.S. economy, namely, the potential to develop and deploy more advanced Internet services that either require IPv6 or run much more efficiently on it. Unfortunately, the rate and scope of market penetration by these new Internet services are difficult to predict. As they become more prevalent globally, the burden on U.S. companies that still emphasize IPv4 while trying to also migrate to IPv6 applications will steadily increase. [ 258 ] Conversely, premature migration to a new generation standard with the high transition costs typical of complex standards (such as the Internet Protocol) can impose large short-term and even medium-term costs on domestic firms Costs: The costs of transition will be incurred unevenly across the industries and user groups that comprise the Internet supply chain. The timing of these costs and their distribution across stakeholder groups will be affected by the appearance of IPv6-specific applications and the degree to which industry and government efficiently execute industry-led transition strategies. For individual organizations within each user group—corporate, institutional, government, and individuals—the transition costs will vary widely. For example, independent users, comprised of home users and small businesses, will likely incur virtually no cost to move to IPv6 as they would gain IPv6 enablement over time without additional testing and installation costs. [ 259 ] Medium-sized businesses, on the other hand, will likely incur the largest relative increase in IT spending to transition to IPv6. The majority of these costs will be related to the core networking operations and staff, the size of which does not increase proportionally to the size of an organization. The magnitude of costs for medium-sized businesses will be slightly less than large organizations, but their annual revenues are significantly lower. Therefore, the costs for medium-sized businesses relative to sales will be much higher. Industry indicates that over the next four or five years the vast majority of network hardware, operating systems, and network-enabled software packages (e.g., databases, email, etc.) will be sold with IPv6 capabilities as users upgrade or replace worn out hardware and software. As a result, users may not incur significant additional hardware/software costs to acquire IPv6-capable IT systems. However, having the capability to run IPv6-based applications is significantly different from having access to and running actual applications. To do so will require the emergence of applications and the “turning on” or “enabling” of the hardware and software. This will require re-training of IT staff. Thus, the Task Force’s research and analysis indicate that labor (training) costs will constitute the majority of the total extraordinary costs of upgrading to IPv6 for users. Benefits: As a technical matter, IPv6 has advantages over IPv4. Over time, the technical advantages of the new protocol will likely produce several types of benefits. There are, however, significant disagreements among stakeholders about the timing and magnitude of those potential benefits, as well as their distribution among providers and users. The most frequently cited infrastructure benefit from the adoption of IPv6 is a vast increase in available addresses for people and machines that need to be connected. Demand for such addresses will likely increase as more and more of the world’s population requests Internet access. The situation may become critical if the projected markets emerge for in-home devices (e.g., “smart appliances,” entertainment systems, voice/video over IP) that need to be accessible from outside the home via the Internet. Although there is considerable disagreement about whether, to what extent, and at what pace, such demand will develop, IPv6 would provide the address space to accommodate any level of demand which emerges. Emerging market applications, especially devices that are globally addressable so that they can be remotely accessed and controlled via the Internet, represent a potentially important application of IPv6 addresses. Further, automobile components or subsystems, refrigerators, cameras, home computers, and other home appliances could be assigned IP addresses, linked together on home networks, and connected to the Internet. Home owners could control such devices remotely, and automobile and appliance manufacturers, for example, could offer remote service and support packages. Wireless sensor networks and machine-to-machine communications will eventually lead to the proliferation of devices that will connect to the Internet. Additional benefits of remote access are the potential increased life expectancies of large ticket items such as automobiles and appliances (durable goods) and an associated decrease in service/repair costs. For example, RTI estimated that a one percent increase in life expectancy and one percent decrease in service costs for automobiles and appliances would yield approximately $3 billion dollars in economic benefits. Network efficiency benefits flowing from adoption of IPv6 could also be significant. Many of the benefits hinge on removing and/or changing the management of NATs, firewalls, and middleboxes, because they currently disrupt certain types of end-to-end connections. A NAT-enabled firewall presents a small number of public addresses to the Internet and, therefore, conserves limited address capacity (a problem under IPv4), while using private IP addresses for all the personal computers behind it. Participating stakeholders indicated that application vendors allocate significant labor resources to design or redesign their products so that they will work through NAT boxes. Some experts have stated that this work could stifle innovation by diverting time away from other infrastructure and application R&D activities and by increasing the complexity of new applications. Security Implications: Over the long term, adoption and use of IPv6 by government, the private sector, and the Internet as a whole may produce security benefits. In the short term, implementation on any new communications protocol, such as IPv6, will likely increase security threats to networks and users. The greatest potential security benefits of IPv6, moreover, appear to depend on the development and implementation of security mechanisms and paradigms significantly different than those commonly employed in today’s networks and largely independent of the particular Internet protocol (e.g., IPv6 or IPv4). Additionally, the transition mechanisms that will be employed during the lengthy migration from IPv4 to IPv6 will likely present their own security concerns and challenges. Effective and secure migration to IPv6 will require careful testing and evaluation, deployment guidance and standards, and development of IPv6-aware security hardware, policies, and processes. Adoption of IPv6 may also necessitate the replacement of existing perimeter security architectures with end-to-end architectures. That will require research and development expenditures for new security management technologies and mechanisms. Consequently, the potential security benefits of IPv6 in the longer term must be balanced against (1) what might be considerable development costs to complete the design and development of these new models, (2) potential increased risks entailed by deploying such models incrementally in existing operational networks, and (3) the ability to deploy security improvements (such as IPsec) without deploying IPv6. In the near- to mid-term, carefully integrated security planning, IPv6-specific security development, and security testing should precede any organization’s decision to deploy new IPv6 technologies operationally so as to ensure the security and stability of both the new IPv6 resources and the existing resources that they may interact with. Failure to do so could easily result in degrading the security posture of the organization’s existing IT systems. Deployment Strategies/Options: The Task Force believes that, consistent with existing laws and within current resources, the federal government has four major roles with respect to deployment of IPv6: • continue to monitor and analyze technological and market trends in global IPv6 infrastructure and applications; • conduct research on IPv6 infratechnologies and facilitate industry standardization processes; • support industry with performance/behavior test methods and test beds, as needed; and • deploy and enable IPv6 to meet internal government IT needs, after adequate planning. Industry expressed a range of views on government roles in facilitating deployment of IPv6. In general, infrastructure vendors and users were more enthusiastic about government involvement than were application vendors and ISPs. Moreover, infrastructure vendors and users differed in the type of activities and technical areas where they believed government should be involved. Additionally, most stakeholders mentioned government research and test beds related to development of scalable end-to-end security models and quality of service mechanisms. Regarding adoption, many stakeholders specifically suggested that government support in standards and protocol development, along with compliance and interoperability testing, should be provided through existing industry standards bodies such as the IETF and existing test beds such as Moonv6. With respect to procurement, the federal government is a major market for IT systems, including Internet applications. It can thus provide an initial market of substantial size, which will both demonstrate to the rest of the economy the value of IPv6 applications and provide data on the most cost-effective strategies for transitioning to the new protocol. Any federal government IPv6 initiative,
however, must include careful planning and both procurement
and deployment strategies to be effective (the previous
experiment with federal procurement policy toward Internet
protocols, GOSIP, only mandated procurement and, as
a result, failed). Moreover, given that industry has
led efforts to date on IPv6 research and development,
standardization, and deployment, a government procurement
and deployment policy will only be successful if government
coordinates with industry and provides internal technical
expertise to assist government agencies to develop and
implement transition strategies.
FOOTNOTES [256]
GAO IPv6 Report, supra note 43, at 31. |
||||||||