Hadoop Behind Metamarkets' Data Science-as-a-Service
Business Intelligence startup Metamarkets aims to take "BI in the cloud" one step further.
As a part of the third episode of InformationWeek's Valley View, Metamarkets was one of the two cloud-based business intelligence startups that gave their elevator pitch to me and Fritz Nelson (see the embedded video below, which includes a demo of the online service).
Compared to the other cloud-based business intelligence startup that was on the same episode of Valley View (see Chart.io Seeks To Shake-Up BI), Metamarkets' unique selling proposition is its ability to crunch enormous volumes of data by relying on two key performance-oriented components in its stack: in-memory technology and Hadoop.
"We work with the raw data, so we bring in terabytes--close to a pedabyte of data in the last year--and Hadoop is part of our solution," said Metamarkets CEO Michael Driscoll. "Obviously, it's open source, and we run the Hadoop processing as the first step in our data processing."
Unlike Chart.io, whose security message claims that the data always resides with the customers, the only way Metamarkets can act on terabytes, let alone pedabytes, of data is to take custody of that data. Driscoll claims that his customers' data is safe with him--safe enough to earn the praises and support of the European Union.
For companies whose data sets become too big and unwieldy for them to manage at scale on their own, letting someone else, like Metamarkets, worry about how to administer it within the context of Hadoop could make more sense. As Driscoll points out, the sorts of data sets that Metamarkets manages often take many more servers than just one to crunch all that data into a meaningful dashboard in a reasonable amount of time. Metamarkets distributes that data across hundreds of servers to achieve scale and speed.
Driscoll, who says that he likes to charge the price of a data scientist (something organizations might otherwise have, in addition to the bunches of servers needed to sort through terabytes of data), said the company's service starts at around $10,000 per month and automatically crunches the data for you.
About the only claim that we don't buy is when Driscoll says that your business intelligence solution should be co-located with your data. By that thinking, if all the data you're crunching is in the cloud (let's say you're using Google Analytics to generate metrics for your Web site), then so too should be your BI solution. We're not so sure about that logic. But we did like the buzzwords that Driscoll used to describe what BI solutions do: "Slice, Dice, Visualize, and Analyze."
Attend the March 28 edition of Valley View and you could win! For each episode of Valley View, one lucky member of our live online audience wins some cool tech gear (for example, we've already given away a bunch of Amazon's Kindle Fires). We announce the winners and the prizes at the end of each show. But to win, you first have to register for the drawing. Or better yet, attend in person at our San Francisco headquarters and meet the editors and guests! If you want to be a member of our live studio audience, just send an email to one of the Valley View hosts: David Berlind or Fritz Nelson.
Informationweek.com run-of-site player, used to publish article embedded videos via DCT. The same ads will be served on this player regardless of embed location.
6 Tools to Protect Big DataMost IT teams have their conventional databases covered in terms of security and business continuity. But as we enter the era of big data, Hadoop, and NoSQL, protection schemes need to evolve. In fact, big data could drive the next big security strategy shift.
Big Data Brings Big Security ProblemsWhy should big data be more difficult to secure? In a word, variety. But the business wonít wait to use it to predict customer behavior, find correlations across disparate data sources, predict fraud or financial risk, and more.