Symons said last week that the first version of the technology will focus on unstructured data such as text files, spreadsheets and PowerPoint presentations, as well as semistructured files such as e-mail.

Brian Babineau, an analyst at Enterprise Strategy Group Inc. in Milford, Mass., said EMC has all the pieces to create a data classification product but must now knit them together.

EMC already has a partner agreement with Arkivio Inc., a Mountain View, Calif., company that makes a data classification tool. But an EMC spokesman said the deal is unrelated to the company's internal development plans.

StoredIQ Corp., another data classification vendor, next week plans to announce a partnership agreement with EMC that's also unrelated to the latter's product plans, said Dave Davenport, CEO of the Austin-based start-up.