Comments on: Kickfire kicks off http://www.dbms2.com/2008/04/18/kickfire-kicks-off/ Choices in data management and analysis Fri, 08 Aug 2008 00:45:11 +0000 http://wordpress.org/?v=2.5.1 By: Curt Monash http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-91527 Curt Monash Thu, 24 Jul 2008 05:08:20 +0000 http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-91527 Thanks, Robert. I should email Rick and nudge him to answer us directly. :) Best, CAM Thanks, Robert.

I should email Rick and nudge him to answer us directly. :)

Best,

CAM

]]> By: Robert David http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-91525 Robert David Thu, 24 Jul 2008 03:35:50 +0000 http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-91525 You wrote: "*Somebody – perhaps adman extraordinaire Rick Bennett? — may want to check my memory on this, but I think Oracle’s famed “Gentlemen, start your snails” ad in the early 1990s was about PC World tests, not TPCs. Oracle also had an ad about WW1-style planes nosediving, but I don’t think those referenced TPCs either." I remember meeting Rick Bennett at his home-based office on the Peninsula when he was working on those WW1-plane ads. I'm pretty sure they were TPCs. -RAD You wrote:
“*Somebody – perhaps adman extraordinaire Rick Bennett? — may want to check my memory on this, but I think Oracle’s famed “Gentlemen, start your snails” ad in the early 1990s was about PC World tests, not TPCs. Oracle also had an ad about WW1-style planes nosediving, but I don’t think those referenced TPCs either.”

I remember meeting Rick Bennett at his home-based office on the Peninsula when he was working on those WW1-plane ads. I’m pretty sure they were TPCs.

-RAD

]]> By: Curt Monash http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-82291 Curt Monash Sat, 19 Apr 2008 21:43:48 +0000 http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-82291 Hi Steve, Should I assume box manufacture is outsourced to one of the (probably Asian) usual suspects, the exact identity of whom is undisclosed? As for the TPCs, I suspect that chip/hardware guys get more excited about benchmarks than dyed-in-the-wool software guys do. :) But seriously, a 2X price/performance improvement over the runner-up alternative isn't obviously "disruptive". What's disruptive is the data warehouse specialist vendors as a group radically outperforming the general-purpose ones. Best, CAM Hi Steve,

Should I assume box manufacture is outsourced to one of the (probably Asian) usual suspects, the exact identity of whom is undisclosed?

As for the TPCs, I suspect that chip/hardware guys get more excited about benchmarks than dyed-in-the-wool software guys do. :) But seriously, a 2X price/performance improvement over the runner-up alternative isn’t obviously “disruptive”. What’s disruptive is the data warehouse specialist vendors as a group radically outperforming the general-purpose ones.

Best,

CAM

]]> By: Steve Dille http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-82214 Steve Dille Fri, 18 Apr 2008 16:16:12 +0000 http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-82214 Thanks for the post Curt. Here are the clarifications you are looking for on some of the implementation and packaging details. I know we covered a lot in the one hour briefing. The Kickfire Database Appliance is an out of the box appliance that includes the commodity CPU hardware, disk and Linux + MySQL 5.1 and the Kickfire storage engine software + our SQL Chip. It is 2 RU and larger versions are 3 RU. Each appliance contains what we call the Base Server Module (which runs Linux and MySQL): This has Dual Quad-Core Xeon CPUs and 16 GB RAM. This is connected via a PCI cable to our Query Processing Module which contains the SQL Chip and various amounts of RAM depending on the model (up to 256GB today). Remember also, that Kickfire both compresses data and operates on compressed data (not something that all appliances and analytic databases do). So what is actually in memory is effectively much more user data. And because of the architecture, our chip operates on data in memory without the bottleneck of registers (you mentioned this above). Storage is included with the appliance as well (8x either 37GB, 73GB or 146GB SAS drives). You also can connect your own external storage to Kickfire (not something all appliances support.) Also, we want to be clear about our announcement and emphasis on TPC-H. This is very exciting technology with disruptive implications on the costs of database query processing. With Kickfire and MySQL, open source data warehousing class performance is now possible in a load-and-go appliance. We just announced the Beta version of our appliance is now available. The TPC-H results were done to demonstrate the power we bring in a small, cost-effective package. Our announcement was done to get the word out to attract innovative clients to our Beta program. General availability is not that far away. We are quite busy. We will stay in touch with you Curt. Thanks. Thanks for the post Curt. Here are the clarifications you are looking for on some of the implementation and packaging details. I know we covered a lot in the one hour briefing.

The Kickfire Database Appliance is an out of the box appliance that includes the commodity CPU hardware, disk and Linux + MySQL 5.1 and the Kickfire storage engine software + our SQL Chip. It is 2 RU and larger versions are 3 RU.

Each appliance contains what we call the Base Server Module (which runs Linux and MySQL): This has Dual Quad-Core Xeon CPUs and 16 GB RAM. This is connected via a PCI cable to our Query Processing Module which contains the SQL Chip and various amounts of RAM depending on the model (up to 256GB today). Remember also, that Kickfire both compresses data and operates on compressed data (not something that all appliances and analytic databases do). So what is actually in memory is effectively much more user data. And because of the architecture, our chip operates on data in memory without the bottleneck of registers (you mentioned this above).

Storage is included with the appliance as well (8x either 37GB, 73GB or 146GB SAS drives). You also can connect your own external storage to Kickfire (not something all appliances support.)

Also, we want to be clear about our announcement and emphasis on TPC-H. This is very exciting technology with disruptive implications on the costs of database query processing. With Kickfire and MySQL, open source data warehousing class performance is now possible in a load-and-go appliance. We just announced the Beta version of our appliance is now available. The TPC-H results were done to demonstrate the power we bring in a small, cost-effective package. Our announcement was done to get the word out to attract innovative clients to our Beta program. General availability is not that far away. We are quite busy.

We will stay in touch with you Curt. Thanks.

]]> By: Daniel Weinreb http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-82203 Daniel Weinreb Fri, 18 Apr 2008 13:17:07 +0000 http://www.dbms2.com/2008/04/18/kickfire-kicks-off/#comment-82203 I'm not sure I see why it makes a big difference whether the chip is custom silicon or an FGPA. Or whether it's three chips or one chip or four chips. The real question is how much it costs and how well it performs. People have been trying to come up with special hardware accelerators for database systems and file systems for quite a long time. So far it doesn't seem to have yielded big results in the marketplace, but past results are not a good predictor of future performance: maybe they've really got a winning technology here. The TPC-H result perhaps is something they're making a big fuss about because it's the main, or only, way they can demonstrate good price-performance. And, of course, real customer testimonials are nice. Their list of VC's is impressive and their management has good credentials. This looks worth keeping an eye on! I’m not sure I see why it makes a big difference whether the chip is custom silicon or an FGPA. Or whether it’s three chips or one chip or four chips. The real question is how much it costs and how well it performs. People have been trying to come up with special hardware accelerators for database systems and file systems for quite a long time. So far it doesn’t seem to have yielded big results in the marketplace, but past results are not a good predictor of future performance: maybe they’ve really got a winning technology here. The TPC-H result perhaps is something they’re making a big fuss about because it’s the main, or only, way they can demonstrate good price-performance.

And, of course, real customer testimonials are nice. Their list of VC’s is impressive and their management has good credentials. This looks worth keeping an eye on!

]]>