Archivas

Newspaper CTO starts his own company

April 29, 2004

2 Min Read
Network Computing logo

Say what you will about the media: Startup Archivas Inc. owes its life to a newspaper.

The Waltham, Mass., startup's founder, Andres Rodriguez, was CTO of The New York Times when he learned that all the news thats fit to print© was tough to fit into a digital archive.

"Nothing was adequate,” he says of technology tools available at the time. “We wanted to do a midrange level 5 RAID, but it was too expensive, so we outsourced it to ProQuest [a service provider that stores digital content]. That means we lost real access to our own data.”

Rodriguez was fed up enough to take action. He assembled a team of engineers and came up with his own solution: an object-based file system for digital content. The Archivas system stores and retrieves files and metadata as objects rather than files. Users write policies for retention, authentication, data shredding, and protection to meet compliance standards, based on the metadata.

Now Archivas's product is in beta at several sites, including NASA's Goddard Space Flight Center in Greenbelt, Md. Goddard is using the software to store atmospheric data retrieved from satellites. Archivas's finished product is set to ship in September.Archivas is set to emerge from stealth Monday, but it got funding of $6 million last May from North Bridge Venture Partners and Polaris Venture Partners.

Archivas already faces formidable odds. Between Rodriguez's time at the Times and the start of the new company, EMC Corp. (NYSE: EMC) released its Centera archiving system to handle fixed content (see EMC Preps Centera). “EMC Centera has defined this space,” Rodriguez admits. And it's clearly tough to be the new guy when the other choice is an established market behemoth.

Other storage systems, such as Isilon Systems' IsilonIQ, also use a cluster file system similar to Archivas's to handle large files such as digital film and medical records (see Isilon Intros Clustered Storage Systems).

Undaunted, Rodriguez thinks Archivas can make it, for several reasons. First is price. Rodriguez says his online archive will cost a penny a megabyte, or one half the cost of Centera.

Rodriguez also maintains that Archivas’s object-based software can survive the inevitable hardware platform changes that come with storing long-term data. Some new regulations maintain that certain records must be kept for decades.“Over thirty years, you’re going to be replacing a hardware tier maybe six times,” Rodriguez says. “If you distribute data through one of our clusters, you let the cluster do its thing for thirty years.”

The Archivas software is certified to run on servers from IBM Corp. (NYSE: IBM) and Penguin Computing Inc., and Rodriguez expects certification from Dell Computer Corp. (Nasdaq: DELL) and Hewlett-Packard Co. (NYSE: HPQ) soon.

— Dave Raffo, Senior Editor, Byte and Switch

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox
More Insights