Open Source Datacenter Conference 2013

Published 2013-04-29 by Jochen Lillich

Recently, I attended the Open Source Datacenter Conference, short OSDC, in Nuremberg. This is a short review of my experience.

The three-day conference is hosted by Netways, an IT services company specialised in Open Source Software, just like ourselves.

I arrived on Tuesday night at about 20:30. Dinner was arranged so that even late arrivals could still get something to eat. Even better was that I found five ex-colleagues back from the good old WEB.DE days gathered around a table. A great start to a great conference.

To be exact, the conference had already started in the morning with workshops. But since I didn’t attend them, I can’t say anything about these.

As the conference programme was structured into the tracks “Cloud & Big Data”, “DevOps and Methods” and “Infrastructure Services”, it covered three of the hottest topics in IT at the moment.

On Wednesday, I attended these talks:*“2000 databases later”: A great look behind the scenes at Booking.com, especially their huge MySQL installation. Kris Köhntopp is a brilliant IT operations expert and his opening session, both informative and entertaining, set the bar to a height that IMHO no subsequent speaker was able to reach.

  • “Introduction to NoSQL with Couchbase 2.0”: CouchBase promises to be an in-place replacement for Memcached which we’re using heavily for freistilbox. But its lack of redundancy is a problem, so I was looking forward to Tugdual Grall’s talk. I left it confident that CouchBase is worth doing a proof of concept installation.
  • “Petabyte storage with Ceph”: As storage technology is another essential topic at freistil IT, attending Martin Gerhard Loschwitz’s talk was a no-brainer. Ceph is getting more interesting every time I hear about it but I don’t think we’ll be using it anytime soon.
  • “Configuration management and Linux packages”: The approach of distributing service configurations by auto-generating and deploying config packages certainly is better than doing it manually. But since, compared for example with Chef, it solves only a small part of the system management problem, Schlomo Schapiro could not convince me that this is a way to go in 2013.
  • Jan Doberstein finished the first conference day with “The truth is in the logs”. His overview of useful log management tools confirmed that we need to get Logstash running ASAP.

In the evening, we left the hotel together for dinner and drinks at the “Indabahn” bar. I spent most of the time chatting with my ex-colleagues (when they weren’t occupied with playing Ingress) about what’s going on in our jobs and lives.

After some good night’s sleep (note to self: make sure to switch off the AC first thing after getting into hotel rooms in the future), I started refreshed into the second half of the conference:*“lxc@libvirt”: We’re already using LXC for freistilbox and I hoped to perhaps catch a few new tidbits. But not only did Erkan Yanar only scratch the surface, his way of presenting also seemed to me unstructured and improvised, so I left his session disappointed.

  • “PostgreSQL in 2013”: While we’re using MySQL almost exclusively, PostgreSQL is gaining more and more traction (for example as the main RDBMS at Heroku) and we’ll need to run it in production for the Chef v11 release. So Michael Renner’s overview over recent developments in the PostgreSQL world provided a good way for me to get up-to-date.
  • “Introduction into Hadoop”: Olivier Renault’s session was another opportunity to learn more about a topic I’ve got no personal experience with.

In the evening, I left OSDC 2013 with a lot of new ideas, many of them spawned by simple side remarks of speakers or experiences shared by other participants. That’s what makes OSDC for me worth attending every time, even if I’m already familiar with many of the topics presented there. I’m going to propose a session myself again next year.

Since there were always two talks at a time, I could not attend all the sessions I had been interested in. Sometimes, I also gave preference to the “hallway track”, ie. talking to other conference attendees outside. I’m going to complete my knowledge intake as soon as the session video recordings are available.

The only downside for me was the long train ride to and from Nuremberg. This will change next year when I’ll be able to fly directly to Berlin where the conference will take place from 2014 on. See you there!

PS: There are just too many interesting IT conferences and DevOps meetups for our small team to attend! Help us!

Previous

Index

Next