On 08/22/2011 03:25 PM, Karsten Wade wrote:
I'll likely end up answering stuff a few times, but that's OK.

Part of this is that I had two ideas at the same time - one to do an
open source conference, the other to call it CUDCon. Clearly it's an
evolving beast. Onward with replies ...

On Mon, Aug 22, 2011 at 01:46:51PM -0700, Robyn Bergeron wrote:

So my gut feelings on this:

* Having this at the *same time* as FUDCon would be less than optimal,  
as we've seen in the past with having FUDCon co-lo'd with JUDCon or  
Summit, or even FUDCon at the same time as other events (fosdem, etc.).  
People have to choose from one or the other, and can't make the most of  
their time at either event.  Even FADs run on the same days as  
$variousconferences wind up having people come in and out and not be  
able to get anything done.
* Having it before/after puts a lot of folks in the position of having  
to take off a full week from work, or else having to choose one or the  
other.
So perhaps my thinking on this is a few years out of date, considering
the alignment of thinking that says, "No, duh!"

The thinking is, bring in a *new* set of people who wouldn't be at
FUDCon.
I understand that. My thinking is, that those people won't want to come to a place where they're feeling like they're going to have Fedora constantly being "advertised."

Don't put up a competing track, merge. All the Fedora-focused cloud
talks end up being done "within" CUDCon. Think of it, if you will, as
a tag saying, "This talk is interesting to Fedora & cloud," i.e., more
than one audience.
Yes, but all of the non-Fedora cloud stuff winds up being in that track as well, and it becomes something that a bunch of us have to choose from. And it's the folks who have been busting ass to make Cloud actually successful in Fedora - we're the ones who wind up getting the most screwed. Because we care about Fedora, and we care about Cloud. It's one thing to have a day-long track - it's entirely something different to expand it to something large where we're obligated to essentially be at one event or another.

      
* I fear that despite having it under "different branding" - even having  
it at the same time as a Red Hat-sponsored "event" will give it the  
illusion that it's going to be very Red Hat-focused - possibly  
disenfranchising other folks from coming.
I'll be honest, I think (as with the original FUDCon), the CUDCon as a
brand is good enough it is worth doing battle with the "Yet another
*UDCon from Red Hat."

But as I said at the start, and is clear in the response, there are 3
points here.

1. Is it useful to have an open source cloud conference?
Yes.
2. Can it be called CUDCon or does it need a more neutral name? But
   what about the humor value? (CUDSummit.org is available, too, but
   loses the -con air.)
While I think it's cute, I think a more neutral name and more neutral venue would make it more neutral.
3. Should it be paired with FUDCon?

Bigger conferences pair stuff all the time and it works. I'm just
wondering if CUDCon is the thing that grows FUDCon to be bigger-er so
it doesn't suffer from the brain-drainage problem.
Yes, bigger conferences pair stuff all the time.  But it's usually the "vendor-neutral" conference that is the main event, and "branded" parts of that may run before or after. Take KVM Forum and Linuxcon for example - obviously LinuxCon was the main event, and KVM Forum was an add on.  But they weren't simultaneous, people weren't forced to choose - they went early, and stayed later. Same goes for things like Build a Cloud Day preceding events like SCALE, OLF, etc.

I don't want FUDCon or CUDCon to be the solution to "fix" the other problem. They should both be able to stand on their own as viable events. As it is, people are *fully booked* when they come to FUDCon with things they already have to solve/hack on. I would say at least a good half of the people active on this list and in meetings, myself included, don't do the Cloud SIG as part of "their job" - it's purely volunteer.  And there are other folks for whom Cloud is their primary job, and maybe even "Cloud in Fedora" is part of their job, but are far more active in Fedora than just that angle - gholms and ke4qqq are the primary examples here. For people like us, and I don't want this to come off as 'all about me' since we all know my actual technical contribution is really really low ;) - our ability to effectively participate in both parts of the conference is basically shot. 

      
I really like the idea of having a Cloud Developer/End user-focused  
event, giving devs the opportunity to work amongst each other and find  
common ground, ways to work together, and getting to hear about the gaps  
that end-users are experiencing -- but I feel like having it be more  
independent, and getting some key folks in from other communities to  
help drive it and put it together, would make it much more  
unbiased-appearing.
OK, let's keep that as a very serious option, that I just take this to
a stand-alone, new plan and discussion, then start pulling people in
from various related communities.

We did have what was more or less a "full track," so to speak, of  
cloud-stuff at FUDCon this year, and that worked very well, even though  
it was somewhat Fedora-focused.

My other major concern is "what happens if it's super successful" (oh  
noes!) - how do we manage that with the limited Fedora budget that we  
have, or where are we getting money to sponsor what could potentially be  
another 100 folks showing up - as far as a "fudpub option", having  
enough rooms at a hotel at a point when FUDCon itself has blocked off  
only enough rooms for a Fedora audience, etc.
I'm using this open thinking process to bake a plan I'll be presenting
to platform and cloud product and marketing folks at Red Hat for
funding -- regardless of FUDCon connection. Do I think the FUDCon
connection makes it an easier or harder sell for those folks? Not sure
yet ...

So my go-plan for Milan is to scrape by with what I can find for
funding (very little) but prove the model, and use that to justify
actual budget for Blacksburg from $sponsors. Naturally, I'd start
looking for sponsorship first at home ...

Generally, I think it would be far better off as an independent event,  
particularly if we want it to be an Independent Event - we can't say  
that we want it to be for everyone and that it's not Fedora-focused, but  
still want to leverage the fact that Fedora is onsite. I really feel  
like it's one or the other, but doing both I think causes a big  
distraction for FUDCon itself, and ties itself in a way to Red Hat  
branding that really makes it not independent, no matter how much we say  
otherwise.  Even calling itself an *UDCon is essentially reusing names  
that are given to other Red Hat conferences, which I suspect probably  
would just give people the impression that it's going to be very RHT  
focused.

Sorry to be all Negative-Nancy.  Like I said, I like the idea - but I  
don't want the idea, or FUDCon, to suffer - I think it could be very  
successful as something that is more independent, both in terms of  
attendees and output done.
You'll note in the proposal that one of the values of connecting with
a FUDCon is the chance to interrelate with the very developers who
work on both important upstreams and important integration points
(Fedora and RHEL/EPEL.) That's a value and focus to start.

I'm feeling like that is something I *can* deliver on by pairing with
FUDCon.

I can't deliver that equivalent value without pairing an open source
cloud conference with another conference.
Do we really need that value? Is the value of having 10+ Big Name open source cloud projects not enough?

If anything, I'd argue that the most value would NOT be in the OS layer - the most value would be in synchronizing with virtualization projects, and with other management pieces (Puppet, Chef, Zenoss, etc.). For the most part, the cloud stuff out there is abstracted away from having a direct relationship with the OS folks by virtue of the fact that most of it is used to orchestrate virtualization bits.  Even if we did try and deliver that value via FUDCon, most cloud folks are in the know enough to know that the only people present are going to be Fedora/RHEL and KVM-focused, and unless we get folks from Ubuntu and Suse and Microsoft and Xen and VMware to show up, the "value" isn't going to be very comprehensive for them to be compelling.

I see there is enough there-there to bring together some of the open
source cloud efforts to a conference, but what brings out the kernel
and packaging and release engineering folks from the various Linux
distros?

If we can get a conference that *starts* alongside FUDCon, there is no
reason it can't run alongside other distro-specific conferences.

And that brings us back around to the saw's edge that I think FUDCon
is at. It doesn't have to grow in size, but if it is going to, one way
is to figure out how to embrace the shared communities of interest.

Anyway, I'm not stuck on any idea other than the same one we all agree
needs doing, 'neutral-ground open source cloud conference'. Well, I'm
a bit stuck in not taking ourselves too seriously, which is why I want
to be all clever with the CUDonyms.

- Karsten
_______________________________________________ cloud mailing list cloud@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/cloud