“Nothing aggravates me more than seeing companies throw away data,” says Andrew Joiner, General Manager of Emerging Technologies and Marketing for Autonomy. “Smart companies are monitizing their data through analysis.”
So why are companies throwing away Pbytes of data that could provide vital insights into their businesses? The main reason, Joiner told SiliconAngle CEO John Furrier and Wikibon CEO Dave Vellante in the Cube at HP Discover 2012 in Frankfurt, Germany, is that just building the induces for multiple Pbytes of data, and particularly for mixed data types, takes huge amounts of processing power and days or weeks in a traditional infrastructure, in which the data and processing are separated.
“The hole grail is to put the analytics as close to storage, and as close to bare metal on the storage, as possible.” That, says Brian Wyse, VP of Information Governance for Autonomy, is what Autonomy’s Idol does. Idol is a NoSQL database that uses a pure mathematical model rather than a word match algorithm. It runs on the data storage server to build induces and run analysis on multi-Pbytes of unstructured data. It is 40-times faster than a traditional system. Building an index for a 2 Pbyte unstructured database that might take a typical traditional infrastructure a week can be done in less than five minutes with Idol.
Integrate that with HP’s other major Big Data analytics acquisition, Vertica, which is very good at handling very large structured databases, and you have a system that can tame the largest Big Data databases, Wyse says. That is exactly what HP has done, and now it is embedding the Idol/Vertica system in the heart of a list of Big Data products that do anything from analyzing huge medical databases to predicting which HP products in which client environments may fail in the next month, allowing HP to provide proactive to its customers. “This is unique to HP,” Wyse said.
Twitter is candy for Idol, Joiner said. “For instance we did an analysis of Twitter feeds for McDonalds. Ninety percent of the messages were directions – ‘I’m a half mile from the McDonalds’, ‘How do I get to that McDonalds,’ that sort of thing. The other 10% were about getting cold fries. That analysis told the company which franchises were under-performing.”
They did an analysis for the producers of Madagascar to judge the reactions of children to the movie. Most of the data was positive, “but we discovered this outlier where the children were running screaming from the theater.” It turned out that the projectionist in that theater had run a movie trailer advertising a horror film before the start of Madagascar by accident.
This kind of analysis, Joiner says, will revolutionize search. “Soon you will no longer be typing search words into Google and getting pages of responses. Mobile devices will take you directly to the information you need in the real world – directions, movie show times, whatever you are interested in at that moment.” That will be powered by the kind of Big Data analysis Autonomy does.
“Meg Whitman has made three big bets as CEO– on cloud, security, and information,” Joiner said. “Autonomy is at the heart of that information strategy. Over the next year you will see continued services alignment and storage innovation, where they are embedding us in the information layer across the product architecture.”