Last week Teradata announced version 15.0 that has much better integration with Hadoop. Teradata really gets what is happening in the data management space. Not only do they now have Aster as part of the Teradata family, they also work closely with Hortonworks and are doing a good job of integrating Hadoop. Teradata’s UDA model is a recognition that one size does not fit all. Their ability to connect well to other databases like Oracle with products like Unity Source Link (USL) makes Teradata a very strong contender to be the centre of Gartner’s Logical Data Warehouse. One very powerful thing that Teradata does is push the processing to the data source rather than pull data into Teradata itself. This is something I strongly believe in and something that Simba has been pursuing for some time now. The SQL Engine built into the SimbaEngine ODBC and JDBC SDK is purpose built to be fully componentized. This means that we can break up the query processing of any SQL statement so that if, for example, the underlying data source supports indexing but does not support joins, then the Simba SQL Engine can let the underlying data source do the indexing while Simba will do the joins. This Collaborative Query Execution (CQE) capability of Simba is something that Teradata is doing at the database level with UDA. Teradata’s QueryGrid technology is what allows Teradata to push processing to the underlying data source whether it is Hadoop or Oracle. Alex Woodie at Datanami wrote a good summary entitled “Teradata Makes Data Warehouse More Hadoop-ish“. I really like what Teradata is doing in supporting a heterogeneous data environment. The world of data is growing exponentially and it is clear that there will be a proliferation of data sources. Making them all work together and work together well is very important.