[K12OSN] OT: Apple Imac lab

Peter Scheie peter at scheie.homedns.org
Sun Mar 1 17:17:57 UTC 2009

Ryan Collins wrote:
> Peter Scheie wrote:
>> Not to put too fine a point on it, but you should let the powers that 
>> be know that the acquisition cost is a minor part of the TCO.  The 
>> major part is in supporting the machines, keeping them running, etc.  
>> With LTSP, you mostly support one machine, the server, with a little 
>> overhead for the clients.  If they get 20 Macs, make sure everyone, 
>> and I mean everyone, knows the cost of supporting them will be 20 
>> times the cost of the LTSP lab.  Linux, done properly, is cheaper in 
>> the short run and MUCH cheaper in the long run.
> TCO for 24 machines will basically be the same as a lab of thin clients, 
> especially if you purchase a copy of Apple Remote Desktop. From one 
> machine you can update the lab, install software, change settings, etc.
No, it's not.  TCO is not just a measure of updates and the effor they require. 
  All those fat clients have several more variables by definition than do thin 
clients, which must all be managed; the cost difference increases as the 
machines age and hardware becomes more likely to fail; performance varies as 
some clients are replace sooner than others; consistency quickly falls apart, 
driving up management costs.   I agree that video editing is probably not 
something one would want to do on thin clients.  My point is that the people who 
say 'We want to do video editing' have to understand the the short and long-term 
costs of such a setup are potentially exorbitant, which seems questionable for 
an application that will be used, what? Maybe two hours per week?  That's less 
than 7% of the time kids are in school (assuming 30 hours per week). And that 
assumes that all students use this app, which isn't likely.  A technological 
approach that costs more than twice as much up front, more than that over five 
years (what kind of shape will those Macs be in 5 year? 7 years?) for something 
that is used less than 10% of the time by a minority of students seems like a 
ridiculous idea.

I think Burke arguably had the best idea, with some caveats: Create a small Mac 
lab for the video editing, but have it piggyback on the LTSP server for 
authentication and home dirs.  LTSP lab(s) should be the default configuration 
throughout the school since it will provide 90% of what students need; and use a 
*small* Mac lab for such a specialized app.  And make it clear how and why that 
lab costs more and provides less than the rest of the computing system (LTSP). 
Frankly, as a parent, I'd rather have the money spent on a(nother) teacher's aide.


More information about the K12OSN mailing list