Kickfire: Released for all the right reasons, but with the worst timing?
Kickfire's hardware-accelerated MySQL appliance blows the doors off monster Oracle database servers in equivalent TPC-H benchmarks in a tiny fraction of the cost, space, storage and power usage. Last week, during their press briefings the future of the company looked bright indeed. There's only one problem -- Oracle is about to own MySQL.
Last week, startup Kickfire briefed analysts and press on the official launch of their first product, the Kickfire MySQL Appliance. I had a chance to talk with Karl Van Den Bergh, Kickfire's VP of Marketing and Alliances, about the product's technology roadmap as well as the partnerships they were going to be announcing at the MySQL Conference and Expo this week. One of the partnerships they announced on monday was a co-marketing agreement with Sun Microsystems, MySQL's owner, to sell the product to Sun's customers.
Click on the "Read the rest of this entry" link below for more.
The partnership announcement, excepting all other major industry news this week, would have seemed to spell success and a great future for the company, not to mention as a great way for Kickfire and MySQL to get into the burgeoning data warehousing market for databases under 10 Terabytes, which represents a $5 billion dollar a year business according to recent figures released by IDC. Up until this week, MySQL data warehousing was expected to be a $500M market per year in and of its own.
Also Read: Kickfire -- MySQL Data Box for the Rest of Us
I first started following Kickfire around the same time last year, when the company made its debut during the last MySQL Expo. The technology is indeed impressive -- if you've got a decision support application with a SQL database under 3TB in size, there isn't a hardware and software platform in that can match it on price and performance. To put it in perspective, imagine getting the same horsepower out of a $15,000 budget compact car that you would get out of a $150,000 exotic sports car -- with 1/10th the fuel consumption and taking up 1/10th of the parking space.
Kickfire is able to crank out those huge TPC-H numbers due to it's Query Processing Module -- a specially designed processor with dedicated cache (64GB RAM worth) combined with altered MySQL code that allows it to perform database operations much faster than a commodity x86 server -- in some cases, hundreds of times faster that would otherwise require a system that costs ten times as much.
But with Oracle purchasing Sun along with MySQL, presumably in order to "buy the books" and Sun's MySQL customer base, it's not hard to predict how well Kickfire's $35,000 hardware-accelerated Linux box for will fare with Oracle's sales team when it is competing with the company's flagship Oracle 11g DBMS, which starts at $10,000 per licensed CPU, and runs on Sun's -- soon Oracle's -- SPARC and Opteron-based Solaris servers.
While reptiles are known to eat their own offspring, Larry Ellison probably isn't going to be so hot on cannibalizing Oracle 11g and Oracle BI sales with Kickfires that cost 1/10th of the price and partner integrates with Open Source competition from Talend, PentaHo and Jaspersoft. You can make a bet that Larry is going to make it awfully difficult for Kickfire to get the support and cooperation it needs from the company, and will do his best to weasel out of the deal, even if it means ending up in court.
Well, what should Kickfire do? For starters, I'd start looking at porting the SQL processor integration to other Open Source databases, such as PostgreSQL, as well as forming relationships with companies like Monty Program AB, led my MySQL founder Monty Widenius, who is working on an independent Open Source fork of MySQL. MariaDB. Heck, I might even consider extending Monty and his team a generous employment offer, and giving MariaDB some real sponsorship and Larry Ellison the middle finger.
Is Kickfire doomed to sudden infant startup syndrome at the hands of Oracle? Will it become a takeover target by one of Oracle's competitors? Or should it simply switch gears, a la Wile E. Coyote, and go to Plan Monty AB? Talk Back and Let Me Know.