Cloud Data Analytics on Ulitzer

Greenplum Software is stepping out of the giant data warehouse business and into something a little smaller and comfortable with the general availability this week of its spanking new Single-Node Edition, a free, potentially viral, data mart version of its database.

It has all Greenplum's Postgres-based big-time parallel analytic capabilities; it's just hardware-restricted.

Greenplum sees it giving more people the data analysis they crave and enterprises a way to collect a lot more of their otherwise dispersed or silo'd information.

Users can also become a distributed node on Greenplum's months-old Enterprise Data Cloud, which offers centralized management, data discovery and data sharing across databases.

Production use of the free widgetry is restricted to an x86 server with two CPU sockets and unlimited CPU cores or a single virtual machine using a maximum eight virtual CPU cores. There's no cap on storage; it can run into the 10s of terabytes but the company figures they won't pass a half-terabyte.

Greenplum has overcome its row bias; it can do as column-oriented processing too.

For those who want it there's a paid support option and naturally an upgrade path to multi-node deployment.

The company says the Single-Node Edition taps into a growing trend - "a new class of elite data analysts working outside the enterprise data warehouse" with huge volumes of mostly raw rather than clean data, use cases a clumsy data warehouse can't satisfy.

Greenplum claims to forging a path toward a distributed, private cloud environment. Its cloud mojo lets customers provision and manage hundreds of databases on a private cloud infrastructure.

Greenplum claims its business is growing faster than its proprietary rivals Netezza and Teradata. The company has been weaning itself off of its dependence on Sun to sell its widgetry in the face of Sun's downturn and its prospective acquisition by Oracle.

It says Sun now represents less than 10% of its sales; it's leaning more on HP, Dell, EMC, Cisco and IBM.

It also claims Oracle's Exadata data warehouse appliance isn't competitive with its standard up-market fare since it's challenged at 10TB and above.