Proposal: Fedora at Home Project
jspaleta at gmail.com
Mon Mar 10 18:45:42 UTC 2008
On Mon, Mar 10, 2008 at 10:32 AM, Bryan Che <bche at redhat.com> wrote:
> We are doing most of this work upstream in the Condor community. So,
> the development work is not Fedora-branded.
> Do you mean a custom Fedora spin? I'd love eventually to have this as a
> first-boot option on a standard Fedora distribution where users can
> opt-in to donating CPU cycles.
That sort of discussion is a ways out I think, but its probably
appropriate as a vision statement in your SIGs page.
> To clarify: if we host the live scheduler at Fedora, would we still
> brand this as a Fedora project even if we support having other platforms
> donate CPU cycles?
An actual scheduler service that runs on Fedora project
infrastructure... yes I would support branding it as a Fedora project.
Since I've no idea what is needed to actual host a scheduler, you'll
have to have a discussion with infrastructure concerning required
resources and then bring a summary of that discussion back to fab.
I'd want to know what the resource burn for a scheduler is expected to
be. If we are doing a scientific task initially, do we have to commit
to storing a large data or results set on Fedora infrastructure as
well as the scheduler?
More information about the fedora-advisory-board