Road to IP v.6 may be long and winding
While no one doubts that the transition to Internet Protocol Version 6 eventually will happen, the challenges in moving to the new protocols could be as great as the opportunities.
RESTON, Va.'The Coalition Summit for IPv6 is bringing together more than 500 evangelists and seekers of knowledge about the new version of the Internet Protocols, the set of rules governing computer communications that promise improved security, mobility and quality of service in our voice, video and data networks.
But the challenges in moving to IPv6 could be as great as the opport1unities.
'IPv6 sounds good, but it has been accepted at about the same level as the metric system,' said Bill Kine, a product manager for Spirent Communications Inc. of Calabasas, Calif.
The label 'IPv6-compliant' is common on networking hardware and software but means little in an environment where specifications for the protocols are developing so rapidly that vendors can't keep up, Kine said. And the added overhead of processing IPv6 can degrade performance on equipment not optimized for the new protocols.
No one attending the summit doubts that the move to the next generation of Internet Protocols eventually will happen. Europe and the Pacific Rim already are far ahead of the United States in its adoption, forced by the scarcity of IPv4 addresses in those regions and the growing use of mobile technology. The question is when will it come?
Richard Beaird, the State Department's senior deputy coordinator for information and communications policy, said IPv6 will be the underpinning of the next generation global network that will combine the reliability and security of the public switched telephone networks with the flexibility and economy of the Internet.
This merger will take place over the next 10 years, Beaird said, and international standards-making bodies such as the International Telecommunications Union (ITU) are defining its parameters.
China, Japan and Korea are playing major roles in creating these definitions. Beaird said U.S. policy on next-generation networking is that ITU recommendations should not create market barriers for U.S. companies.
But Kine said even in Europe and the Pacific Rim, IPv6 networks still are quite small, usually no more than pilot projects and test beds. 'IPv6 real world experience is really quite limited,' he added.
Because standards are still evolving, compatibility and interoperability still are major hurdles. And because IPv4 networks are expected to be around for another decade or more, clumsy dual stack, translation or tunneling technologies will be required to accommodate both kinds of traffic. This will interfere with promised advantages of operating a native IPv6 network.
Although Rep. Tom Davis (R-Va.) earlier Tuesday called for government to take a leadership role in moving to IPv6, Kine said market demand will be the surest way of bringing about the transition.
'With apologies to Rep. Davis, I believe that a killer application will drive adoption much faster than any government regulation,' Kine said. Everyone has his own idea of what that killer app will be; Kine said it will be mobile computing, as millions of new devices such as cellular phones become IP enabled.
For Microsoft Corp., the killer app will actually be an operating system. Jawad Khaki, Microsoft's vice president of networking and communications, said the next version of Windows, code named 'Longhorn,' would drive adoption of the new protocols.
Longhorn was released to hardware vendors for product development in April, and a client beta release is scheduled for this summer. General availability is estimated in late 2006.
Longhorn will be fully IPv6-compatible, with IPv6 the preferred transport protocol. All connections will be attempted first with IPv6, and IPv4 will be used only as a fallback.
Larry Roberts, CEO of Anagram Inc. of Redwood City, Calif., and the leader of the Advanced Research Projects Agency team that created the Arpanet 40 years ago, sees quality of service as the essential enabler for IPv6 adoption.
The inability of IPv4 to easily handle QOS designations makes the Internet an unreliable carrier for time-sensitive traffic such as streaming voice and video. Roberts has proposed a QOS signaling scheme that would use a 16-bit block in the IPv6 packet header to indicate priority.
Roberts said increased network capacity can allow in-line QOS signaling, and memory on IT devices has improved to the point that data can be maintained on the state of each traffic flow, making real quality of service provisioning possible. 'This allows us to start real-time video and voice intermixed with data,' he said.