May is conference month and the memorable event–for all the ideas and progress over the past year–for me was HBaseCon.

HBase achieved a milestone earlier on this year with the release of 1.0. Note that the rolling upgrade feature from 0.96 makes this and future upgrade a very reasonable proposition.

The recent highlight for HBase has to be the trend for using HBase as a persistence layer instead of writing to HDFS directly. Trafodion and Kylin are notable examples for this approach.

The other major move is Apache Phoenix’s introduction of a query server to host its query processor. This opens the way for a thinner JDBC driver and, more importantly, an ODBC driver. Might we see something this year? Together with the adoption of Calcite for the query parsing and optimization, Phoenix brings HBase into the realm of the SQL-on-Hadoop engines. But seeing that Phoenix is focused yet on read-access, it is probably premature.

I was sorely disappointed that I couldn’t make the SQL-on-HBase smackdown panel at the end of the day. The videos aren’t up yet but I’ll be the first to report it when it gets posted. A mini set of interviews (including of yours truly) by SiliconAngle is available for the while. Enjoy!