HP Starts to Craft API Strategy Around Autonomy

Although the acquisition of Autonomy by Hewlett-Packard may go down as one of the most questionable financial transactions in the history of IT, the core search and indexing technology that Autonomy developed is emerging as the lynchpin of a new set of API services that HP intends to expose to application developers. Robert Youngjohns, general manager of the HP Autonomy business unit, says HP is in the process of making sure the Intelligent Data Operating Layer (IDOL) addresses some of the challenges developers are incurring in the era of Big Data. To help developers address those issues, HP Autonomy plans to make the core IDOL functions available as discrete APIs that developers can invoke either on premises or in the cloud. Rather than forcing organizations to adopt IDOL in its entirety, the HP Autonomy API strategy is part of a larger effort to embed IDOL within a broad array of HP and third-party applications, Youngjohns says. HP, for example, already uses IDOL as the foundation for delivering a variety of applications, ranging from backup and recovery and data governance to Web content management. As application evolve in the Big Data era, Youngjohns says, there is a need for indexing and search services that can be readily invoked, so each developer doesn't have to redevelop these services. As part of that effort, HP Autonomy will develop an object store capability to make it easier for developers to persistently access HP Autonomy services, he says. While there’s a lot of interest in raw data these days, as a practical matter, developers will want the ability to invoke data sets that have already been created using raw data. Those data sets could be stored in HP Autonomy in a way that would turn them in reusable services running on the HP Cloud or elsewhere that could be invoked via an API. To make that vision a reality, HP Autonomy is also creating a cloud-based set of development tools that will make it easier for developers to build applications that invoke the HP Autonomy APIs. The end result, says Youngjohns, is a way to build next-generation machine learning applications that would rival IBM Watson without requiring massive investments in manual services to actually implement. Autonomy may not have been highly valued at the time HP acquired it, but the fundamental technology that HP gained via the acquisition is at the heart of a larger data management strategy that includes Hadoop and HP Vertica. Obviously, it will take HP several years to hone strategy in a way that appeals to developers. But once it does, chances are the IDOL component of the HP Autonomy acquisition will prove to be a foundational component of an overall effort to build new classes of data-driven applications.

Michael Vizard