Hewlett-Packard took a lot of flak when it decided to buy Autonomy for $10 billion, but if big data is a worthwhile investment, perhaps HP’s alleged extravagance will pay off. In a phone call Thursday afternoon, Autonomy Chief Marketing Officer Nicole Eagan explained to me why she thinks HP is in a good position to compete with IBM, Oracle and the rest of its peers when it comes to analytics.
Autonomy has been painted primarily as an enterprise search company, but that might not be a fair characterization. According to Eagan, the company’s IDOL (Intelligent Data Operating Platform) is designed to work with any type of data, and lets Autonomy’s products do everything from, yes, search to analysis of videos to sentiment analysis of social media data. In fact, she said, Autonomy can be a good replacement even for Hadoop, especially for more-complex workloads such as natural-language processing. Whereas Hadoop itself is just the data-processing engine that requires a natural-language processing layer to be bolted on, Eagan said Autonomy can do the same job by itself and more completely because it’s designed to decipher intent even from poorly expressed language.
When combined with HP’s Vertica analytic database, Eagan thinks HP has a fairly complete data story, at least as it relates to what we typically call big data. Primarily, Autonomy can process and analyze unstructured data while Vertica handles structured data. However, she noted, combining unstructured information with structured data stored in Vertica can result in some very good social graphs.
Eagan also commented on Autonomy’s 31-petabyte cloud database full of unstructured data for customers, saying they do far more with it than just search.
As high as she is on her own company, Eagan is equally skeptical about Oracle’s decision to purchase Autonomy competitor Endeca yesterday. “On the one hand, it really goes to validate the market for unstructured data,” she said. But, in her view Endeca still isn’t much more than keyword-based search with some additional functionality. Unlike Autonomy, which is based entirely on mathematical algorithms that let it adapt to new data types and workloads, she says Endeca is limited in scope.
Although, it should be noted, Endeca does have a product designed specifically for e-commerce that Oracle already plans to incorporate with existing software targeting that market. And Oracle CEO Larry Ellison had some choice words for both HP and Autonomy when the two companies announced their deal in August.
If Eagan isn’t overselling Autonomy, and if HP uses the IP it’s acquiring to its full potential, it looks as if HP actually will have a story to tell in big data. That will be good for competition, and probably for a lot more mud-slinging among IT heavyweights trying to stake out their big data reputations.