[arin-ppml] Is this more desired than aTransferPolicy? Needinput
Jo Rhett
jrhett at svcolo.com
Thu Nov 20 02:24:35 EST 2008
On Nov 19, 2008, at 11:00 AM, Ted Mittelstaedt wrote:
> Exactly. Most LARGE companies. As in, most cash-rich, forklift
> upgrade all PC"s every 3 years so they are all the same, with a
> large MS site licese, companies.
>
> Why exactly are we worried about these companies? They have the
> cash to go to IPv6. And if they are doing rolled-out OS upgrades
> then they are running XP or Vista and can turn on IPv6 quite
> easily by just rolling out a config change to all seats.
Because these are the same companies who will need 14-18 man-years of
effort just to evaluate IPv6 implementation in their business-critical
software.
>> The comparative cost of doing the application conformance
>> tests alone,
>> assuming all of their switches, routers and firewalls are
>> already IPv6
>> compliant, dwarfs that cost.
>
> I simply disagree. You have to do the same amount of application
> conformance
> testing when your going to a new Windows OS. Consider the millions of
> bucks these companies are paying Microsoft every year for their MS
> site licenses. IPv6 rollouts require NO licensing fees!
Have you ever done conformance testing for a mid-large business which
is not an ISP? Given your response, I can see the answer is clearly no.
> I don't see that learning
> about IPv6 is any tougher than the transition from MS Domain-style
> networking to MS Active Directory.
Good data point, let's examine that. Most companies took 10 years to
make that change, and had 3 or more failed rollout attempts before
they succeeded. Estimated cost to a 200-person company I was
consulting for at the time was in excess of 20 million dollars. Yes,
that cost was $100k per desktop machine.
Again, they received a perceived advantage by making this change.
There is no real or perceived advantage (today) to deploying IPv6.
They aren't going to even start this project until they are forced to.
> Your average network admin wasn't considering that they would have to
> go to 64-bit machines and replace all their server hardware when they
> went to the new version of Exchange, but they managed it.
>
> Your average software developer is paid to know these things. If
> they don't then customers will stop buying their software and go
> to a competitors software that does know better.
>
> I think you are simply not giving credibility to the abilities of
> the average network admin. I'd rather operate from the position that
> these folks will be able to manage this transition just as they
> managed
> every other transition.
I spent 20 years working with network admins, and consulting mid-large
companies at the highest levels on technology. I know what skill
levels are available in the administrator level, and I know intimately
what their costs are for major migrations. Your attempt to map a
major, touch *EVERYTHING* networking change to a simple server
replacement suggests that you don't grasp this problem at all.
Please take a step back and consider the real effects. Go talk to a
network administrator at a mid-size company who is not a technology
provider, but depends on technology.
And I don't mean to be rude, but it's clear that your beliefs outweigh
any actual experience to the contrary, so I'm going to stop replying.
More information about the ARIN-PPML
mailing list