This year’s recent Cassandra Summit EU represented the great growth taking place in the community.  Back in March of 2013, I brought my family over to the UK to help establish a formal business presence and the support team for DataStax in Europe.  I hadn’t anticipated encountering so many deployments from such a wide range of companies and organizations.  At the summit this year, there were small startups, large enterprises, banks, gaming companies, music streaming services, utilities, mobile apps, and so forth.

The day before the general Summit sessions, there was a full day of training.  I attended Patrick McFadin’s data modeling track which was filled to capacity as we brought in additional tables and chairs.  As always, Patrick was both entertaining and informative.  The content has matured over time to add a bit of formalism to the approach.  While every data model and query pattern will be unique, teams can benefit from experience of others with the general principles and design templates.  There are always different training opportunities available through DataStax but there is also a good set of presentations that Patrick did on data modeling found here: http://wiki.apache.org/cassandra/DataModel

The day of the keynotes and general sessions was alive with several hundred attendees.  It was great to see familiar faces as well as names that I had only seen on Jira, Twitter, the mailing list, and through interactions via DataStax.  During Billy’s keynote it was interesting to see the companies represented and sharing the stage.  British Gas’ connected homes initiative is interesting to me as an example of ingesting and analyzing large amounts of sensor data to help, among other things, proactively predict furnace failure.

I enjoyed seeing content from Jonathan Halliday and his co-presenter Rebecca Summonds.  They spoke about the challenges and lessons learned in indexing geo tagged twitter data and storing that in Cassandra such that they could efficiently query it.  It was also good to have a conversation afterwards with Jonathan and Ben Hood who has been working hard to implement a Go driver for Cassandra, found here: https://github.com/gocql/gocql.  We had a good conversation about client drivers and what functionality is and could be exposed.

Alexander Dejanovski from Chronopost spoke about lessons learned trying to educate and reorient colleagues to use a non-relational data store within a company.  Even though impending necessity motivates the transition, investment in effective education and reorientation can smooth the path and shorten the time to production.  A good example of this was working with Demonware/Activision as they converted a legacy sharded mysql setup to Cassandra and did their testing leading up to the launch of Call of Duty.  I was involved with some initial conversations with Séan and the team there and it became obvious that these guys did their homework.  They put everything through its paces, learning everything they could about best practices and how to efficiently use the system both on the development side as well as operationally.  Because the launch is so important, they tested and tested and then did more testing.  They tested at their scale many, many times and did a variety of analyses on different characteristics they found.  It was gratifying to see how quiet the launch was under the load.

For the presentations from Apple, it took a little while to recover from the astonishment of the scale they run Cassandra at.  However it was interesting to hear a deeper story in different contexts.  Sankalp went into some of the lessons they’ve learned as well as some of the contributions that they have made to the project given their scale.  Richard Low gave a presentation about being a “Cassandra Doctor” at Apple and shared some great detective stories describing how they tracked down difficult problems.

I often enjoy the presentations, but the best part is catching up in hallway conversations with what everyone is working on.  For example, it was great seeing Andy Cobley, the Scotland’s patron saint of Cassandra.  In the evening he did a great lightning talk on research into genes and product proteins and how to harness that information.  It was nice to finally meet Robert Stupp who wrote, among other things, the user-defined function support coming in Cassandra 3.0.  Former Hailo architect and London Cassandra community godfather Dave Gardner was also there in his new role as an infrastructure team lead at Apple.

In all, it was a well organized event filled to capacity.  If you ever have the opportunity and are using or considering Cassandra, the Summit is worth the time and travel sacrifice.  You get a glimpse of how others are (and are not) using Cassandra.  You get to meet people in the community that are working on similar problems and learn from their experiences, both from presentations and perhaps more importantly in informal conversations.  I look forward to seeing the recorded presentation videos, but being at the event provides rare opportunities.