6Sense: Generating New Possibilities in the New Internet.
Produced by: IPv6 Summit, Inc.

Getting Experience Using IPv6
By Stan Barber
Vice President, NTT Communications

Stan Barber
NTT Communications

As we all know, IPv6 has not been widely adopted in the US for a myriad of reasons, which have been recounted in various articles in 6Sense over the years. There are many who now expect the US Government OMB Mandate to drive more extensive IPv6 deployment. In turn, this would increase the adoption of the protocol, creating a dual-stack network with IPv6 operating alongside the current IPv4 transport. A few continue to insist that IPv6 is not useful enough to implement and would prefer to see other alternatives. What those alternatives might be remain diverse and uncoordinated, so no widely supported consensus on an alternative has arisen. As long as there is no consensus, IPv6 remains the only viable successor to IPv4.

However, IPv6 (like IPv4) is not a panacea. While the standards work for IPv6 started well over a decade ago, and the basic protocol has been stable for almost a decade (RFC 2460 was published in 1998), there remains concern about multihoming and the address allocation plan currently being used by registries (ARIN, RIPE, APNIC, etc.). The IETF and others are still working on refinements to some of the new capabilities in IPv6. This work may make others uneasy concerning its real readiness. These activities center on Mobility, QoS, use of the Flow label, handling of multi-addressable hosts and such. None of these concerns make IPv6 unusable for production traffic today, but this is important work that will allow IPv6 networks to realize capabilities that were not really available with IPv4. These issues were highlighted in the recently circulated draft profile for IPv6 in the US Government. However, NIST acknowledges that the IPv6 standards are stable and "operationally viable commercial implementations are becoming available." See page ES-1 of that report available at http://www.antd.nist.gov/usgv6-v1-draft.pdf.

The key question for any particular organization that remains is what to do about bringing IPv6 into that enterprise's network. In last month's 6Sense, Karl Sill (echoing the 2005 GAO report on IPv6) cautioned everyone that lack of action about IPv6 does not mean that IPv6 is not active on your network. Some action is required even if a particular organization does not plan to do a full-blown implementation of IPv6 right away. However, for those that do plan to bring IPv6 into the enterprise network, there is really no reason to wait on the availability of the technology. It has been in the router software for some time now, and with the release of Microsoft's latest desktop operating system, Vista, any upgraded desktop will run IPv6 out of the box. The IPv6 Summit later this month is a great opportunity to get up to speed on what is available and where to get it (if it is not already available in the hardware and software on your network today).

Many groups and resources (both free and for-hire) have emerged to help an organization implement IPv6 in its enterprise's network. Determining which to use will depend on many things that are specific to the needs of the organization looking to add IPv6 to its enterprise network. I would strongly recommend doing extensive research on both the needs of the organization and the potential resources prior to making any major investment. Perhaps the most important decision to make when choosing to use a group or resource in an implementation project is how well it can work with the existing team. This is not an IPv6-specific requirement, but a pragmatic reality for managing change in any organization. Again, the IPv6 Summit is a good opportunity to gather information on some of these groups and resources.

I would also encourage experimentation with IPv6 as soon as possible. Implementing a test bed network does not involve a major investment, and will provide an environment to both learn about IPv6 and develop strategies for deployment. To foster this, a test bed should reflect as much of the architecture of the current enterprise network as possible, and should certainly include those aspects that are critical and unique to the way in which the network is used by the organization today. Extending the test bed network with a connection to the IPv6-based Internet will enhance that learning experience even further and provide a platform to refine deployment strategies. Commercial off-the shelf (COTS) IPv6-based Internet access is now available from a few commercial providers in the US (NTT Communications is one. See www.us.ntt.net/ipv6). There also remains at least one free option from Freenet6 (www.go6.net).

NTT Communications continues to be a firm supporter of IPv6, as well as the premier worldwide provider of commercial IPv6 transit. While at the IPv6 Summit, NTT Communications will be demonstrating a use for IPv6 that is currently undergoing tests in Japan for providing earthquake alerts a short period of time in advance of the event, so that people can take evasive action before the tremors arrive. Please stop by and learn more about this unique application.