Hewlett-Packard is among the main proponents of creating so-called data lakes in the enterprise to consolidate disparate records for more efficient analysis, but that hasn’t stopped its investment arm from backing an emerging startup called Tamr Inc. that’s trying to sell CIOs on the exact opposite approach. Instead of centralizing the information itself, its namesake software centralizes access.

The platform taps into organizations’ existing installations of popular data stores such as Hadoop and Oracle to map out the contents using homegrown machine learning algorithms calibrated to match related details across multiple sources. The complete records are then assembled into a catalog that is continuously updated with new data and can be searched as a single repository.

That realizes both the stated goals of the data lake model – removing the accessibility problems of disparate silos while allowing for unified processing of records in multiple formats – without any of the main logistical hurdles. Chief among them is the need to move records from upwards of dozens of different systems supporting a large enterprise into a central location  before any analysis can be performed, which can require more effort than the potential advantages justify.

The immense difficulty of pulling that off makes Tamr’s pitch all the more convincing, but it’s not only traditional data lakes that it’s competing with. Another startup called Alation, Inc. exited stealth earlier this year with a similar information cataloging service that boasts value-added business intelligence functionality that its rival has opted to leave for partners to fill in included out of the box.

The new $25.2 million in funding will help Tamr to convince the market that its approach is better by fueling increased marketing and development efforts. The round included the participation of several big names besides HP, including Google Ventures, Thomson Reuters and the venture capital arm of Massachusetts Mutual Life Insurance Company, among others, and bumps its total raised pass the $40 million mark.

Photo via Marius B