Karl Van den Bergh, VP of Marketing & Alliances for the analytic appliance company Kickfire, recently spoke to me about challenges associated with the data warehousing mass market. If you want a detailed 101 on the DW mass market and how Kickfire is currently addressing it with its unique technology (see press release here), listen to this podcast.
Listen to or download the 11:54 podcast below:
---TRANSCRIPT---
Give me a brief background of Kickfire and tell me a little about the new announcement you guys have.
Sure. Kickfire is a company in the analytic appliance space. We were founded a couple of years ago by two successful entrepreneurs, Raj Cherabuddi and Joe Chamdani. They originated back in the Sun where architects and Sparc processor at Sun, left Sun and found another company actually prior to Kickfire called Sanera, successfully sold to McDATA and then founded Kickfire a couple years ago. We have also a very experienced CEO, Bruce Armstrong, who dates all the way back to Teradata. He was one of the original guys there and ended up being the president of the company before handing off to Mark Hurd, and then was a GM at Sybase.
Also, a co-founder of Broadbase, so he's been in the allied application for data warehousing space for quite some time. We're 50 employees are so headquartered here in Santa Clara and backed by some of the best CCs in the valley Accel, Greylock Mayfield and recently Pinnacle. Last year, we launched the company at MySQL Conference. At that time, we announced four TPC-H world records. TPC-H, for those unfamiliar with is the industry benchmark for data warehousing. And we launched the company and broke four world records in price performance and performance on a single node.
We recently raised a Series B of 20 million and spent last year testing the product with earlier adopter customers. So the announcement for this week, we're officially launching our product which is the first MySQL analytic appliance targeted at the data warehousing mass-market and I'll explain a little bit about the mass-market in a little bit. Along with that, we are announcing a number of important partnerships most notably with Sun. So Kickfire today bundles Sun's MySQL database with our appliance so we're essentially reselling MySQL Enterprise with each of our appliances. With this announcement, we are now seeing a reciprocal relationship from Sun so Sun's sales force, particularly the database sales force, will be now incented sale the Kickfire appliance and bring Kickfire into their data warehousing deals.
We also announcing a number of partnerships with System Integrators so we are as a product company very focused on building and selling the best product possible. We are believed though that service is an important component of any successful data warehousing deployment and are therefore announcing a series of partnerships with leading North American system integrators with expertise in data warehousing and MySQL.
We're also announcing a customer, Mamasource, who is a leading Web 20 company with over 2 million members and they have deployed Kickfire for their analytics and seeing 10 to 100x performance improvement.
Now, speaking of data warehousing, can you tell me what makes the data warehousing mass market unique?
Yes, so again, kind of a 101 on data warehousing. Data warehouse is not market where it's been around for a couple of decades now. But it's -- when people collect their data from their transactional systems and a want to analyze that data, they offloaded to a separate system so it doesn't impact the performance of their transactional system and they're able to do sophisticated queries or analytics sometimes called on that data.
So typical data warehousing query will be tell me who are the top 10 customers who I've sold to in the western region over the last three years. And those are generally more complicated queries on larger data volumes as so this market has emerged. Teradata is kind of the pure play leader and I mentioned Bruce, our CEO, was president of that company but that's a very, very, high-end system.
In recent years, there have been end roads from other vendors to target the middle of that market. So from the sort of 1o terabytes, 25 terabytes in size. Teradata and others will target these very large high end of the market, 25 terabytes, 100 terabytes even to have petabyte today, very, very large data sizes. The middle of the market the 10 to 25 range is something that has been a focus for startups -- for more recent startups some who are no longer really startups like Neteeza, who recently went public, has been successful not middle of the market.
The data warehousing mass-market as we define it is those deployments with 10 terabytes and less. And according to IDC, that market is about 75% plus of the actual deployments today. So of all the data warehouses deployed today, three quarters or more are in that mass-market and that's the market that Kickfire is focused on.
So can you tell me what the challenges associated with the data warehousing mass market are and how Kickfire is addressing them?
Sure. So, I mean the market is obviously appealing in that it is -- it forms a large chunk of the number of deployment, 75% plus as I said. In spite of its appeal though, it is challenging and it's why the market has been a difficult market to crack for data warehousing vendors and why most are focused on the high end to the middle end of the market. So number one, this is a market that like any other data warehousing market needs performance. The difference is that they're highly price sensitive. So where as the very large enterprises can afford to spend tens of millions of dollars on Teradata.
The middle of the market can afford even to spend a couple million dollars on a system. And more recently, with the startups that have come to market whose offerings are in hundreds of thousands of dollars, the middle of the market can vary like price point. When we talk the mass-market, hundreds of thousands dollars, millions of dollars is a lot of money. And the mass-market really need something in the low tens of thousands of dollars.
Number two is this mass-market doesn't have the data warehousing expertise that the middle or high end of the market can afford. They can't afford to hire a full-time data warehousing expert DBA. They have also limited IT resources, they just have very few people who are focused on IT operations, most are very operational and business focused.
And lastly, these are early stage data warehousing deployments so they have what's turned an industry very commonly mixed workload environments. And what we mean by that is it's not like a traditional data warehouse where you have historical data and a couple of analysts running complex queries against it but no one else really gets to touch the system.
In the mass-market, these data warehouses are very operational in nature and you get a lot more people hitting the system, this sort of democrazation of data if you will that's happening in the mass-market. So more and more people want to hit the system, the data needs to be refreshed more frequently because again it's more operational in nature. People often run transactions on the system and so will require having a traditionally safe environment. These are also sort of the challenges that have made the data warehousing mass-market a difficult market to crack.
Now, going back Kickfire, what's unique to the company from a technology perspective?
There are a couple of things. Three, let me just say three things. The first is our SQL chip. So this is the world's first SQL chip. We've essentially done for SQL what Nvidia and ATI and others did for graphics. So we've created a chip that natively understands SQL and has allowed us to get tremendous performance efficiency improvement. So one of our chips can replace times tens of CPUs and deliver the same performers but in the single-chip. The second technology component is our column store and column store engine has become popular in recent years in the data warehouse world.
What makes our columns store unique is that first we are fully ACID compliant which is not easy for column store to do which means you can run transactionally safe applications. Second, our column store supports high user concurrency so we can have 100 plus concurrent users on the system and over a SAAS and Active users on the system. And third, our column store, which deploys compression, is also advanced in that we can execute directly on compressed data whereas other column stores would explode the data once it's in memory.
The third component is we run standard MySQL. So our appliance looks like MySQL running on a Linux server and what that means is that we benefit from the skills sets in the marketplace today that support MySQL and Linux. So it's pretty easy to find that whereas other data warehousing vendors have proprietary databases and need to hire a specialist to support your data warehousing solution. We also benefit from MySQL's third-party applications, connectivity to be to MySQL works out-of-the-box with hundreds of third-party applications and because Kickfire essentially looks like MySQL on a Linux server, we automatically benefit from those hundreds of third-party applications and are able to connect to them.
That sounds like great technology. Now, how are customers using that today?
Well, what we've seen to mention -- what we're focused on because we are -- look like MySQL on a Linux server, we are focused on people who are trying to use MySQL to today for data warehousing. And MySQL has come a very popular database from online applications, Web 2.0 companies, e-commerce, Software-as-a-Service, marketing service providers. These are all companies that are using the livestock and MySQL is the "M" in the livestock.
They're trying to use MySQL for analytics. They've used it for their websites, now they want to use it for their analytics to analyze the data that they're collecting on their sites but what they're running into is severe performance issues. So MySQL running on general purpose hardware really struggles when it comes to analytic queries. And what Kickfire does is we allow those customers to swap out their general purpose hardware and plug-in Kickfire, the purpose built hardware from MySQL.
And customers like Mamasource, which we are announcing this week has seen up to 657x performance improvement relative to where they were on their general purpose hardware. So these customers with a couple of points and clicks can take their existing data warehouse running on MySQL on the Juniper hardware port to Kickfire and immediately see 10 to 100 even more performance improvement -- x performance improvements on Kickfire.
Leave a comment