Groovy punts SQL turbocharger
Touts 'real-time' RDB
Groovy Corporation has launched an SQL database acceleration appliance, boasting of extreme speeds.
According to Groovy, its Groovy SQL Switch "was 105 times faster and processed 37 times more orders per minute than the fastest relational database" when running the Dell DVD Store Database Test Suite.
Groovy claims that when performing the test - which models an online DVD store, hence its name - its appliance had a 2.33 millisecond response rate, handily trouncing the "previously fastest performance" of a "well tuned version of MySQL."
The tests also showed Groovy's setup handling 168,845 orders per minute. The plain-vanilla - although "well tuned" - MySQL system managed only 4,614 orders per minute.
Even if these numbers, as is so often true for best-case benchmark reporting, are at the extreme end of the SQL Switch's performance, they indicate that the company stands behind its goal of providing "a truly real-time, live data experience."
In a canned statement, Groovy CEO Joe Ward said: "Mass-market relational databases were never designed for real-time. Many data-driven businesses developed solutions to work around this issue as the value of real-time became more obvious, but these fixes most often resulted only in a near real-time experience and a costly infrastructure."
The Groovy appliance "democratizes real-time data," according to Ward, and provides "the benefits of the live web to any enterprise that wants to make the leap to the next generation of online services."
The SQL Switch appliance is based on Groovy's SQL Switch parallel SQL-processing software and runs on an Intel Server System S7000FC4UR 4U box powered by four six-core Dunnington Xeon 7400 processors.
Intel and Groovy announced their partnership earlier this year. At that time, Narendra Bhandari of Intel said that "We see Intel and Groovy pioneering a substantial lead in high performance scalable, LIVE data management."
Groovy's Ward is, of course, correct when he says that relational databases were designed for a different time and place. They operate on a "query and wait" paradigm. Groovy's raison d'être is to bring relational databases up to the responsiveness required by today's instant-need web implementations by eliminating the "wait" part of that scheme and instead pushing data to the user without the need for a refresh update.
More information on the Groovy SQL Switch can be found here. ®
This sounds like a half arsed data grid to me, with slack kind of continuous query. If I want a data grid I'd use a data grid product, this half way house between a grid and and RDB doesn't seem to support either paradigm very well.
I also agree that people who think SQL == SQL Server should have their geek license revoked!
It has the answer to the query before the database can produce it apparently.
Or maybe it is a fast "intelligent" cache written to run the benchmark very well.
Have you read their PDF ...?
First, the logo at the bottom right corner of the document is a blatant rip-off of the Oracle logo, Psychologically very subtle.
Secondly, it seems to consist of nothing but marketing bullshit that would be very difficult to test in a real-world situation.
Kudos if it actually works, though.