[nSLUG] linux in the (Windows) computer lab

George N. White III gnwiii at gmail.com
Tue Sep 28 12:06:07 ADT 2010


On Mon, Sep 27, 2010 at 11:13 PM, Stephen Gregory <nslug at kernelpanic.ca> wrote:

> On Mon, Sep 27, 2010 at 08:39:21PM -0300, George N. White III wrote:
>
> <snip>
>
> nice writeup.
>
>>
>> With modern hardware, Virtual Appliances provide adequate performance
>> for classroom exercises.
>
> For compute intesive tasks like R you should get near native
> performance too.

Real, as opposed to classroom, tasks in remote sensing tend to be I/O
and memory bound,
which is not ideal for a VM.  Fortunately, the NASA software dates to
the mid-90's and goes to extraordinary lengths to minimize memory
demands so it could run on workstations with 256M RAM (and costing
upwards of $20k!).

> I am a big fan of virtulization.

There are many tasks that don't come close to the resource limits of
modern PC's. With a modern PC and an apt-mirror repo you can whip up a
new VM in just a few minutes.  A VM can't be more reliable than the
host, but there are situations where the VM "protects" the host from
app problems (e.g., crashing the X11 server or runing out of memory or
disk) so in some cases a VM can salvage a system that would otherwise
be unstable.

-- 
George N. White III <aa056 at chebucto.ns.ca>
Head of St. Margarets Bay, Nova Scotia



More information about the nSLUG mailing list