Friday, April 13, 2012

Dumping/Loading schema in Cassandra

This handy command line will dump a schema from Cassandra:
echo -e "use your_keyspace;\r\n show schema;\n" | bin/cassandra-cli -h localhost > mySchema.cdl

I always forget the "-e" on echo, which is why I thought I would blog this.  We've started using ".cdl" as the extension, short for "Cassandra DDL".

Coincidentally, you can then load it with:
bin/cassandra-cli -h localhost -f mySchema.cdl

Hope people find this useful.

Monday, April 2, 2012

Cassandra vs. (CouchDB | MongoDB | RIak | HBase)


Here is why in "Cassandra vs.", it's Cassandra FTW!

Our organization processes thousands of data sources continuously to produce a single consolidated view of the healthcare space.  There are two aspects of this problem that are challenging.  The first is schema management, and the second is processing time.

Creating a flexible RDBMS model to accomodate thousands disparate data sources is difficult, especially as those schemas change over time.  Even given a flexible relational model, to properly access and manipulate data in that model is complicated.  That complexity bleeds into application code and hampers analytics.

Given the volume of data and the frequency of updates, standardizing, indexing, analyzing and processing that data takes days of time across dozens and dozens of machines.   And even with round the clock processing, the business and customer appetites for additional and more current analytics are insatiable.

Trying to scale the RDBMS system vertically through hardware eventually has its limits.  Scaling horizontally through sharding becomes a challenge.  Operations and Maintenance (O&M) is difficult and requires a lot of custom coding to accommodate the partitioning.

We needed a distributed data system that provided:
  • Flexible Schema Management
  • Distributed Processing
  • Easy Administration (to lower O&M costs)
Driven by the need for flexible schemas, we turned to NoSQL.  We considered: MongoDB, CouchDB, HBase, and Riak.   Immediately we set out to see what support each of these had support for "real" map/reduce.  Given the processing we do, we knew we would eventually need support for all of Hadoop's goodness.  This includes extensions like Pig, Hive, and Cascading.

CouchDB dropped out here.  It supports map/reduce, but little or no notable support for Hadoop proper.  MongoDB scored "acceptable", but the Hadoop support was not nearly as evolved as the support in Cassandra.  Datastax actually distributes an enterprise version of Cassandra that fully integrates the Hadoop runtime.   Thus, we left MongoDB for another day and scored HBase's Hadoop support off the charts.

Riak is interesting in that they provide very slick native support for  map/reduce (http://wiki.basho.com/MapReduce.html) via REST, while they also provide a nice bridge from Hadoop.  I must admit.  We were *very* attracted to the REST interface. (which is why we eventually went on to create Virgil for Cassandra)

Left with Riak, HBase and Cassandra, we layered in some non-functional requirements.  First, we needed to be able to get solid third-party support.   Unfortunately, this is where Riak fell out.   Basho provides support for Riak, but Datastax and Cloudera were names we were familiar with.  

NOW -- Down to HBase and Cassandra.  For this comparison, I won't bother re-iterating all the great points from Dominic William's great post.   Given that post and a few others, we decided on Cassandra.

Now, since choosing Cassandra, I can say there are a few other *really* important less tangible considerations.  The first, is the code base.  Cassandra has an extremely clean and well maintained code base.  Jonathan and team do a fantastic job managing the community and the code.  As we adopted NoSQL, the ability to extend the code-base and incorporate our own features has proven invaluable. (e.g. triggers, a REST interface, and server-side wide-row indexing)  

Secondly, the community is phenomenal. That results in timely support, and solid releases on a regular schedule.   They do a great job prioritizing features, accepting contributions, and cranking out features. (They are now releasing ~quarterly)   We've all probably been part of other open source projects where the leadership is lacking, and features and releases are unpredictable, which makes your own release planning difficult.  Kudos to the Cassandra team.