April 21, 2008
"Imitation is the sincerest of flattery."
- Charles Caleb Colton (1780-1832), from his Lacon, Vol. I, published in 1820Welcome to the Data Warehouse Appliance club - another validation of an important, growing market segment
Well, well, well! "Only" eight years after Netezza coined the term and invented the market segment, Teradata today finally officially entered the Data Warehouse Appliance market. Though it's a bit late, and certainly behind a number of other vendors, perhaps today's entry will put an end to Teradata's vacillating over whether they 'invented' the concept or not, were an appliance or not, or whatever. In the past couple of years, it seems Teradata spokespeople have gone out of their way to say their product was simultaneously a data warehouse appliance and absolutely not one - even booking appearances on panels of data warehouse appliance "vendors". Certainly their announcement is another validation that the role of Data Warehouse Appliances is an important and growing one not only in the current market, but for the future as well.
Derivative Marketing and a "Repackaged, Warmed-over" Product?
Teradata is positioning this new product as being, "simple, powerful and cost-effective" - which to our way of thinking sounds much more than a little derivative from Netezza's"Performance, Value and Simplicity", but I'll leave it to the reader to decide if you think so. Our reading of the Teradata announcement sounds like just another larger vendor's "repackaging" alternative to respond to the competition. Like others before them such as IBM and Oracle, it appears that with the 2500 model Teradata has done nothing more than cobble together a collection of elements from the company's model 5500 systems, repackaged and sold as an appliance.
Powerful. Um, How's That Again?
And while anyone who is serious about the appliance segment of the data warehouse market (like Netezza) has focused on delivering systems that can scale to highly complex, enterprise-wide, high performance systems, we think the 2500 will struggle to deliver even modest performance for just 6 TB in a single equipment rack.
While Teradata is quoting just over 6 TB of user capacity per two nodes in this new system, let's remember that they have been advising customers for the past year not to put more than 1.5 TB against each of those same dual-core CPU nodes. Which is it? Is the 2500 underpowered for its 6 TB data capacity per dual-node rack, or has Teradata been advising its model 5500 customers to pay at least 2X too much for their data warehouse systems for the past year?
Time will tell whether Teradata has made other compromises to the 2500 model in an attempt to limit its impact on its flagship products (5500 and the new 5550). Beyond its underpowered nodes, have they sacrificed anything else like workload management or system availability, or even the system's ability to handle highly-interactive, operational applications? As the days and weeks help raise the shroud covering the model 2500 further, we'll know more. For now though, it just feels like "me-too" imitation.
