DataStax Developer Blog

Slides you shouldn’t miss from the Cassandra Summit

By Jonathan Ellis -  June 27, 2013 | 0 Comments

These are my favorite slide decks from the 2013 Cassandra Summit. I say “slide decks” because some great talks really require watching the video; I’ll post about these once we have them uploaded.

Architecture, design, performance

Eventual Consistency != Hopeful Consistency. This one’s at the top because there’s a lot of FUD around about consistency, and Christos Kalantzis absolutely nails it. I’d subtitle this “How I learned to stop worrying and love ConsistencyLevel.ONE.”

My keynote discussed CQL, the state of the art in Cassandra 1.2, and what’s coming in 2.0. I added links to more details on the 1.2 features in this post.

In When Bad Things Happen to Good Data, Jason Brown from Netflix describes the Cassandra write and read paths, along with hinted handoff, read repair, and active repair.

The World’s Next Top Data Model is part three in Patrick McFadin’s data modeling series. Check out parts one and two as well.

Jason Rutherglen’s talk on DataStax Enterprise explains how we integrate Solr with Cassandra and the new features in 3.0.1 and 3.1. Good overview if you’re curious what we’ve been up to on the search front lately.

Extreme Cassandra Optimization by Al Tobey delivers on its title, drawing on Al’s years of experience with Cassandra at Ooyala. He covers filesystems, RAID, hardware choices, leveled vs size-tiered compaction, and much more.

Use cases

Barracuda Networks: Michael Kjellman describes his experience migrating from MySQL to Cassandra 0.8 and beyond, including his famous release day upgrade to 1.2.0. Since moving off of MySQL, Barracuda has grown their dataset by an order of magnitude while cutting latency in half — but this didn’t come for free. Check this out if you’re considering a similar upgrade.

Hailo: Dave Gardner, who runs the Cassandra London Meetup, also moved from MySQL to Cassandra. He talks about some of the nuts and bolts involved, including client choices in Java, PHP, and Go, multi-region clusters, encryption, compression, and more. Tim’s talk explaining Acunu Analytics in detail really needs video to get the most out of.

eBay: Anurag Jambhekar and and Feng Qu discuss eBay’s experience with Cassandra since 2011: hardware choices, performance and scaling, tuning, operations and monitoring, and use cases.

Comcast: Boris Wolf explains the open source cloud message bus, comprising CQS and CNS, queuing and pub/sub services that are compatible with Amazon SQS and SNS, respectively. This joins the Netflix RSS recipe on my list of projects to give someone looking for “more than a toy” Cassandra examples.

BlueMountain Capital: Jake Luciani and Carl Yeksigian gave a very DevOps-oriented talk, covering data model, architecture and production monitoring and performance tuning of Cassandra 1.2.

Just for fun

MariaDB and Cassandra: How MariaDB can map and query Cassandra columnfamilies. Unfortunately this is not yet CQL-ready, but it could become an interesting lightweight alternative to Hive for some users, particular in a mixed MySQL/Cassandra environment.

Cassandra on Raspberry Pi: a 64-node cluster for $3000, and more. Andy Cobley plans to use this with his students — one of whom, not incidentally, was a winner of the first Next Great Data Developer contest.



Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>