Software Tags Data For Automated Security And Information Lifecycle Management

Abrevity has upgraded its FileData Classifier software to better exploit metadata for classifying data for security information lifecycle management (ILM) purposes.

The company's FileData Classifier 2.1 application is a major improvement over its earlier version 1.0, which did only basic classification of the files based on simple attributes such as file names or path names, said Eric Madison, director of marketing and product management for the Cupertino, Calif.-based storage software vendor.

The new version has richer classification capability with the ability to recognize keywords and phrases, said Madison. It is now also able to recognize patterns and proximity data, he said.

"For instance, you could set it to search for a social security number, not the words 'social security,'" he said. "And with proximity recognition, when using an OCR (optical character recognition) document, and someone has written 'social security' or 'SSN' followed by a number, the software knows it might be a Social Security number."

Also new with version 2.1 is the ability to move data to a secure location, or copy it or delete it, based on customer-set automated policies, Madison said. The earlier version of the software depended on third-party data movement software to accomplish the task. "We found out that most small and midsize enterprises didn't have a file movement application," he said.

With FileData Classifier 2.1, each file can have an unlimited numbers of security schemes, said Madison.

The new software allows, for example, IT administrators to look for credit card numbers in order to lock down files with such data without exposing those numbers to the administrators, Madison said. "It can be set up so he can manage the data, but the credit card numbers are extracted from his view," he said. "But the account payable guys can do queries on the credit card numbers. This can be set up across the company, or on a user-by-user permission model."

The software allows the appending of file genealogy, or the actions taken on a file, to the data, Madison said. "Most of the time, if you take some action on a file such as moving or updating it, you overwrite the audit data," he said. "We append the records to leave an audit trail, including when it was accessed and by whom. It becomes an audit trail for compliance. Or, when the data is moved to tape, and then moved back to disk, you don't lose the audit trail."

The same data can be flagged by the software in multiple ways for use with different users, Madison said. "My boss's favorite example is, Rush Limbaugh sees a document, thinks it's a conservative document, and flags it as conservative," he said. "Then Al Franken sees it as a liberal document, and flags it as liberal. Same file, multiple tags."

The FileData Classifier software sits on a Windows-based server to monitor data as it flows. It does not sit on the data path, and can be pulled if needed without impacting the integrity of the original data, Madison said.

Sales for the software go completely through solution providers.

For classifying a single database, the software costs about $7,500 for up to 3 Tbytes of data, Madison said. A manager version of the software, which allows views of multiple databases under a single pane of glass, starts about $20,000 for 3 Tbytes, he said.

The company expects to unveil a plan by a storage hardware vendor to bundle its FileData Classifier software with a disk-and-tape appliance, he said.