Sparky-III will use virtualization (xen). Virtual machines can be
dynamic -- can actually come and go! One big thing to realize, is that
virtual machines can be created and destroyed, booted, shutdown -- all
via remote access.

Normal operation of our traditional services (mail lists, website and
wiki -- maybe a couple other miscellanea) are expected not to need or
use all the system's resources (CPU, memory, disk storage, bandwidth ..
and power, I suppose).

In addition to the very useful prospect of trial implementations of
upgrades and new features, the thing that seems really interesting is
the ability to do experiments.

We can now have a *Kplug Research Department*.

So, how should we manage that. People may think of different research
projects. Some may be found valuable, and adopted as regular services.
Some may produce answers to questions and then cease operations. Some
may need more resources than others. Some may need argu^wdiscussing.

In Kplug tradition, I suggest we should not need anything formal, but
think we might encourage people to suggest experiments, and somehow, the
steering herd will decide and allocate research privileges (and duration
and memory allocations, etc).

I do think some wiki space might serve to collect and massage proposals,
and report things.

All:
 Any mistakes or wrong assumptions above?
 Management-howto suggestions?
 Other comments?

Regards,
..jim

-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-steer

Reply via email to