When the ITU-T IPv6 Group gets together in Geneva next week, one of the things they should discuss is the need for their very existence.
For many decades, those of us with some longevity in the information and communication technology (ICT) field have witnessed recurrent cycles of rampant protocol politics. At its worst, these debilitating cycles occur when the political concerns are not even relevant anymore.
The current outbreak is dubbed IPv6itis, and unfortunately it is at its worst in a venue that should know better -- the ITU-T -- which has spun up two groups that are needlessly consuming international institutional resources that could be applied instead to deal with major infrastructure issues rather than protocol trivia. Two strong arguments and some useful history are provided for a significant re-direction.
Relevancy of the mandate
At the outset, anyone dealing with this subject matter should read Laura DeNardis' book Protocol Politics, published in 2009.
It is almost entirely about IPv6, and reviews in copious but readable detail the controversial history of IPv6 since it emerged in 1994 as a reaction to controversies and politics of the time.
Although the book is oriented for the TCP/IP Internet community oriented, it does establish that IPv6 was a geek response to the Internet Architecture Board's adoption at Kobe in 1992 of the joint ITU-TISO OSI Internet protocol known as CLNP. CLNP was strongly supported by all governments (including the United States) and most of the industry at the time.
When IPv6 was finally adopted, many in industry made it clear they were not going to use the protocol. After 16 years of evangelization, the "father of the Australian Internet," Geoff Huston, who became tired of the endless IPv6 hype, published a definitive set of measurements in 2008 that demonstrated that only 0.4% of the TCP/IP traffic was IPv6. Another set of measurements done in April of this year refined the analysis that indicated a 5% capability of end-to-end IPv6 use currently exists. It is not apparent that anyone has disputed these measurements.Anything that has only captured these low levels of market share after one and a half decades would under most circumstances qualify as a market failure. However, governments -- including via the ITU -- seem obsessed with continuing to drive IPv6 as some kind of panacea with all kinds of political spinoffs represented by the ITU-T IPv6 group. For those who watched the same governments attempt to drive the OSI Internet protocol CLNP for two decades, IPv6 is following an eerily similar path.
Ironically CLNP had a higher relative usage rate at this point than IPv6. As was experienced 20 years ago, this "top down" drive toward a particular protocol solution, notwithstanding the obvious marketplace resistance, is also due to institutional and economic momentum engendered by those who have invested in IPv6. The intent here is not to dump on IPv6, but rather to suggest that the government acting as a marketing agent for specific protocols in this circumstance seems unwise.
The key question is why has IPv6 failed to get traction? DeNardis' book reviews much of the history including the IPv6 concerns from the outset. However, the simple answer today is that IPv6 was designed to meet the operational needs that existed 20 years ago. Conditions have changed dramatically, and the Internet working requirements of 1993 are not those of 2010.
Governments should more appropriately focus on major needs rather than specific protocols. An overview of some of these needs can be found in Dave Meyer's presentation at the recent meeting of the North American Network Operators Group entitled "It's The End Of The World As We Know It" (aka "The New Internet Architecture"). Meyer is one of the long-time Internet protocol doyens who, together with others like Yakov Rekhter, constitute a kind of Supreme Court for routing paradigms.
Several years ago they were brought together by former Cisco and now Juniper Chief Security Architect Barry Raveendran Greene to find a solution to the many profound scaling, routing and security challenges that are important today and going forward. The result was LISP -- the Locator/ID Separation Protocol, which has been perfected through the IETF standards process. It provides a compelling direction that IPv4 or IPv6 lack. Ironically, LISP emulates some features of the OSI IP protocol.
So all of these developments beg the question: why is any group anywhere wrapped up in IPv6 politics? Why is any intergovernmental organization or government agency spending any time on this topic?
The answers lie in both the complexity of the subject matter and the substantial institution and economic inertia that exists. Like the situation 20 years ago, it took years for re-vectoring to occur.
As late as 1992, the U.S. Dept of Commerce was still promoting CLNP, OSI domain names and issuing rules. It was not until 1998 that it discovered TCP/IP and ARPA domain names -- ironically holding discussions in the same meeting room as they had six years previously.
At the same time the FCC, which has responsibility for the public telecom infrastructure, took the wise stance established by the former head of DARPA who was its chief scientist 10 years previously, that the commission would not support specific protocols. Unlike the technologies or marketplace, government policies do not change rapidly, and when they do, it is usually precipitous.
The second argument regarding the need for a significant re-direction concerns the appropriateness of one standards body dealing with allocating or assigning the identifiers of another standards body
The ITU-T IPv6 Group was given what are essentially political tasks:
* To draft a global policy proposal for the reservation of a large IPv6 block, taking into consideration the future needs of developing countries, as outlined in paragraph 23 of C09/29.
* To further study possible methodologies and related implementation mechanisms to ensure 'equitable access' to IPv6 resource by countries.
* To further study the possibility for ITU to become another Internet Registry, and propose policies and procedures for ITU to manage a reserved IPv6 block.
* To further study the feasibility and advisability of implementing the CIR model for those countries who would request national allocations.
* To assist in the implementation of the project called for by Resolution 64, taking into account the needs at regional and national level in terms of capacity building and allocation policies.
* To report to ITU Council 2010.
One obvious question missing from this list is the appropriateness of the ITU-T as one standards body dealing with the implementation of a protocol developed and evolved by the IETF, another independent standards body.
The attempted answer to the question is lame. The assertion that the ITU is participating in the work of the IETF on a virtually non-existent basis certainly does not justify taking over any allocation and assignment of what are plainly the identifiers of another organization without its request or consent.
Furthermore, the IP operations community has a constellation of its own global, regional and national venues, and do not participate in ITU-T forums.
The deployment and use of internetworking protocols and identifiers as well as the evolution of the associated architectures is a highly dynamic, complex and substantially market driven task. ITU-T's IPv6 remit includes tasks that are not only unneeded, but have a strong chance of impeding the very objectives being sought.
With LISP geographical identifiers tacked on the front of existing IPv4, IPv6 or IPv-whatever addresses, the "equitable access" concerns go away. Indeed, from a cybersecurity perspective -- which LISP also facilitates -- making this happen sooner rather than later is worth encouraging.
Disbanding the IPv6 group may be wishful thinking. Political inertia is difficult to overcome. However, it seems worthwhile to reduce the ITU-T IPv6 Group initiative -- if not all governmental IPv6 evangelizing -- to a minimal level and begin syncing with the real-world operational and protocol communities. This includes a transition to LISP.
LISP should diminish the "old protocol politics," but is certain to give rise to its own new political dimensions -- especially relating to cyber/ICT security and involve an array of organizations, including the new UN 15-nation security group.
A trusted implementation of LISP will, among other things, provide a level of attribution and routing security that does not exist in the present architecture. What seems important is to encourage the concept of a diverse institutional ecosystem where no one organization occupies the center of the universe and everyone has a significant stake in the outcome, and where the operator community has maximum flexibility and incentives to implement the transition.
Rutkowski is an expert in international telecommunication technology and law. Over the past 40 years he has served in diverse senior positions in industry, government and academia, including the Geneva-based ITU-T. He is co-author of The ITU in a Changing World and many other legal, technical and historical treatises in the ICT field. See www.ngi.org.
Read more about lan and wan in Network World's LAN & WAN section.
This story, "Disband the ITU-T IPv6 Group" was originally published by Network World.