[AusNOG] IPv6 - What Should an Engineer Address when 'Selling' IPv6 to Executives?

[Paul Brooks]

On 6/03/2013 1:46 PM, Paul Gear wrote:
> On 03/06/2013 12:16 PM, Alastair Johnson wrote:
>> On 3/5/2013 4:58 PM, Mark Newton wrote:
>>>
>>> The network is the easy bit.  If the members of this list had started
>>> low-key activities back in 2011 when APNIC's last-/8 policy commenced,
>>> they'd be done by now, the staff working on it would have received
>>> ample on-the-job training about how it works, and the additional cost
>>> they'd have incurred would have been pretty minimal.
>>>
>>> I'm sure that treating IPv6 as if it's some kind of massive budget-
>>> wrecking major project is part of the reason it hasn't been happening.
>>> Well, it isn't one of those.  It's some planning, then it can go
>>> through the rollout phase largely as business-as-usual.
>>
>> +1 and nicely written, Mark.
>
> Nicely written, but moving away from the question again.  Mark, what are these
> low-key activities that we should have done in 2011 that are easy to sell to
> management?  Feel free point me to something (practical) to RTFM if it's out there,
> but i think Don's overall point that it's hard to cost-justify and hard to get
> moving at the grassroots level in anything more than a test environment is
> reasonably sound.

I think Mark's point was that you *dont* sell it to management. You don't treat it as
a large scale project. You just do little bits of it while you are working on other
stuff in the normal course of business. When management asks you to spin up a new DNS
server, you don't ask management for permission to enable an IPv4 address on it - so
why would you ask permission to enable an IPv6 address on it? 

You define 'IP enabled' as meaning 'IPv6 and IPv4', not 'IPv4 enabled only'. You treat
lack of IPv6 config on an interface as broken - and fix it in the background, just as
you would if you found something else was broken while you were fixing or building
something else.


Paul.