[arin-ppml] Encouraging IPv6 Transition (was: Clarify /29 assignment identification requirement)

[Chris Engel]

If you are actually talking end-user enterprises here as opposed to transit providers then I would forward that the availability/costs of obtaining IPv6 address space is pretty much a non-factor in inhibiting adoption. There are a TON of other factors which do inhibited adoption. Actually obtaining the space would be trivial in comparison to those.  Most of us don't obtain our space through direct assignment anyway, we go through our ISP's/Hosting Companies for it.

Right now, for most end enterprises IPv6 is all pain and no gain. Obviously there are special use cases where that is not true, but for the average end enterprise there is absolutely no business case to deploy it other then to pick up a little advanced experience with it as a hedge against future needs. However, in todays work business/atmosphere most of us are struggling to find time/resources to meet business goals that are due in 10 days not 10 years. Not ideal, but it is what is. It's a very fortunate enterprise these days that has spare resources to get ahead of the game and start playing around with things that may or may not become a priority a few years down the line. If engineers are doing that, in most cases they are doing it on their own time.

I appreciate that the folks involved with ARIN want to speed IPv6 adoption as much as possible as it will resolve the biggest headache that ARIN faces, managing address scarcity. However, I honestly don't believe there is much of anything ARIN can do policy/pricing/procedure wise to significantly effect adoption of IPv6, at least as far as the end enterprise goes. Most of us won't move on it until we are absolutely forced to do so....which likely for many will be a LONG time coming... it'll be an ugly and painful transition but we'll find a way to muddle through it, as we always have....and eventually with market pressures we'll get the tools and solutions we need to in order to support it well.


An interesting anecdote that speaks to where IPv6 still stands on most end enterprises priority list. A few months ago, I attended a local seminar hosted by our preferred firewall/security appliance vendor (Watchguard). During the break, I asked one of the vendors Rep's what the state of IPv6 support was on their latest appliances. I was the only person in the room who mentioned IPv6 at all, and the room included a pretty diverse audience of engineers from both public and private enterprises (local businesses of different sizes, hospitals, schools, police dept's, etc). The rep. stammered for a bit, as he clearly hadn't had anyone else who had raised the issue in recent seminars he had given and then referred the question to one of his colleagues. The colleague proudly proclaimed that they "officially supported it" and then showed me a web page with an "IPv6 Ready" logo/tagline for one of their latest products. The colleague then cautioned me and the rest of those present that under NO circumstances should we actually enable IPv6 support on any of their devices.  According to the vendor,  "officially supporting IPv6" meant that their firewalls/security appliances were capable of ROUTING IPv6 traffic... however, as the rep. explained, the devices Rules Engine and Packet Inspection Filters were incapable of processing IPv6 packets, so enabling IPv6 support on any of those devices would turn the security appliance/firewall into an OPEN ROUTER, completely bypassing all perimeter security controls. Needless to say, no one present was much interesting in experimenting with such a thing.

Now that's anecdotal and I realize it's only one vendor, but I think that speaks alot about where IPv6 still is at in terms of the average end enterprise.






Christopher Engel