SQL in HBase Thursday, January 31, 2013

Salesforce (the company I work for) recently released a SQL layer on top of HBase. Like James Taylor says, "We put the SQL back in the NoSQL"!. SQL is a nice language for asking questions about your data. Having it available with HBase means you need not write code to answer many kinds of questions.

As a performance engineer and one of the early users of phoenix, I am impressed about its performance. It uses several techniques like Parallel scans, filtering on the region servers, aggregate pushdown, hash joins, etc. Even if you don't use phoenix, you can learn quite a bit about HBase performance just by seeing how phoenix goes about executing queries.