There were two very interesting industry announcements today. The first (which I’ll highlight with no false modesty) was our own — the introduction of Greenplum Database – Single-Node Edition. This fully-featured parallel analytical database is entirely free for production use on a server with up to 2 CPU sockets (or 8 virtual CPU cores when running on VMware, Amazon EC2 or other virtualization layer). I’ll have more to say about this later, and you can read why Curt Monash thinks this is a significant and very cool offering.

The second big announcement came from Teradata. This was something they called the “Teradata Enterprise Analytics Cloud” and encompasses rapid mart deployment, sandboxing, and agile analytics in a private cloud model. The industry consensus so far is that this is a response to our Enterprise Data Cloud™ initiative, with a few noting that Teradata must have felt very threatened by it to have thrown together a response so quickly.

That may be true, but nevertheless we welcome Teradata to our world. We know first hand how much pent-up customer demand there is for this kind of solution, and how frustrated customers are with expensive and hard-to-deploy proprietary hardware and all the barriers to agile data analysis in traditional data warehousing environments. Teradata’s embracing of this vision is good for the industry because it forces other vendors to up their game and do more for their customers.

That being said, we can only expect Teradata to change so fast. This is a company that generally charges at least an order-of-magnitude more for its flagship product as its competitors, and espouses an impossible vision of unifying all of the enterprise’s data in a single huge database. So how does this reconcile with this new vision? It doesn’t really, since from a product standpoint there is only one new piece. The ‘Elastic Mart Builder’ is a free portlet for their management tool that helps users carve off chunks within a single monolithic Teradata system. Apart from this it is just a new packaging of their mainframe story — buy a single big Teradata system, and rely on their workload management (aka ‘Active System Management’ - only available in their most expensive flagship product) to divvy up the capacity. Hmmm – sounds awfully like the old Teradata story.

By contrast, Greenplum’s EDC™ takes a much more modern approach -- a distributed data warehousing infrastructure on commodity hardware. Administrators can provision warehouses and marts with a few clicks using low-cost commodity hardware, and manage any number of far-flung servers and databases as a unified distributed infrastructure. Users can share and subscribe to data across marts — easily bridging silos to discover and combine data in powerful new ways. It is about flexibility, simplicity, and power in the hands of analysts.

Finally, the last part of Teradata’s announcement was that users can run the free developer-focused Teradata Express edition on Amazon EC2 for non-production workloads. But to what end? This is just a toy, since users can’t grow it up into anything without buying a mainframe. By contrast, Greenplum’s free Single-Node Edition is anything but — unlimited volume of data (to 10s of TBs) on a server with 8 or more CPU cores, and the ability to dynamically grow these into larger systems or plug them into an EDC to share and discover data.