Building the Cloud: Notes on Apache CloudStack (incubating)
Notes on work that's going on in the Apache CloudStack (incubating) project. Event announcements, progress reports, and more.
ApacheCon North America Wrap-Up
Last week I had the good fortune to attend (and speak at) ApacheCon North America. While the event is still fresh in my head, I wanted to jot down a few notes about the conference and talks I attended.
Despite attending roughly nine billion open source conferences (perhaps a slight exaggeration) over the last few years, this was my first ApacheCon. I had expected a slightly larger event, but it turns out the conference had somewhere in the neighborhood of 300 people, which made for a pretty cozy event.
For my money, events are interesting at 300 people to about 1,500 people. At 300 (give or take) you can interact with a sizable chunk of the attendees, and it usually feels pretty low-key. Events much larger than 1,500 people just seem overwhelming.
Luckily for me, my talk was on Tuesday right after lunch – and it was a talk I’d given a version of previously, so I was able to hone my slides before the talk rather than creating a whole new presentation for ApacheCon. The upshot was I had the opportunity to attend and enjoy quite a few talks.
As I’ve already covered Kirk Kosinski did a few talks on troubleshooting CloudStack. Those were really useful for folks who are using CloudStack currently, or will be soon. Kirk spends quite a lot of time in the “trenches,” and gets to see all sorts of fun and interesting issues.
I also enjoyed Chiradeep’s Software Defined Networking in Apache CloudStack, Patrick McGarry’s Powering CloudStack with Ceph RBD and Sebastien’s DevCloud: A CloudStack SandBox. Basically, most of the “Cloud Crowd” talks were pretty good.
Outside the cloud track, there are a few other talks I really enjoyed that I wanted to call out:
So, Daniel’s topic seemed a bit dry at first and I was not convinced by the session title. He won me over, though, when he admitted he had lost his slides and was making do with some slides he’d thrown together and the willingness to lead the session with a lot of audience participation. He drove the session well, and we had a lot of discussion – worth turning up for, and showed a lot of passion for driving the talk.
Nóirín’s talk was also non-technical, but so desperately needed. Not so much a revelatory topic as something that needs to be reinforced in communities that rely so heavily on text-based communication: emails don’t carry the same information that a face-to-face conversation does. Don’t assume that humor, for example, carries through in text. Realize that you’re talking to people who may not be native English speakers. Being overly terse may not work to your advantage – people may read negativity into your message that isn’t there. (People rarely “see” positivity that’s not there.)
I also really enjoyed Luke Kanies’ session Thursday morning. He was an excellent choice, though non-obvious since Puppet isn’t, you know, an Apache project. I’ll be writing that one up separately.
Of course, the most interesting part of any conference is the Hallway Track. I enjoyed getting to know folks from other Apache projects and spending more time with some of the Apache folks that I know a little but haven’t talked to at length.
One “complaint” I have is that the evening events were stand-and-mingle affairs with no seating. While this is better for quick chats and introductions, I really didn’t get to have any lengthy conversations there. A sit-down dinner for committers/etc. would be good in the future.
Another issue with ApacheCon – that will likely be addressed, as it was brought to the producer’s attention by a number of folks – was with the staggered schedule. Some talks broke at different times, which meant it was difficult to switch tracks if you wanted to see a talk outside your track. For example, my talk on Tuesday ran from 2:45 to 3:45. Two other tracks ran from 2:30 to 3:30. Unless the tracks are wildly different, it’s problematic for attendees because they effectively miss out on many talks simply because they start mid-way through another session.
I’m bummed that I missed the BarCamp day on Sunday and Monday sessions because ApacheCon overlapped with SCALE. (Not that I regret being at SCALE, though – it was, once again, excellent.)
Programming ApacheCon must be a bit difficult, given the diversity and number of projects under the Apache Software Foundation (ASF) umbrella. From the Apache HTTPD Server to Hadoop to Tomcat to CloudStack, it’s a really big tent these days. What’s interesting to Hadoop folks may or may not be interesting to Tomcat users/contributors or HTTPD users/contributors.
Effectively addressing an audience as large and broad as the ASF is difficult. ApacheCon NA managed to skim the surface pretty effectively, but many of the communities present could carry a conference on their own. (And some do.) For future events, I’d like to see a little more effort on cross-community cultivation and cooperation and maybe a little less project-specific focus.
Overall, though – enjoyed ACNA 2013 and am pleased that I attended.