At least, that's what Andres Rodriguez learned during his tenure as CTO for The New York Times. In 1999, Rodriguez headed up a project to digitize and archive its 150 years of content. In the end, "we put up about 10TB, but there was ten times more than that that we wanted to archive."

The New York Times' inability to put all their content online inspired Rodriguez to found Archivas, which this April announced its Archivas Cluster (ArC), a "fixed content repository" that runs on standard Linux hardware, and that can scale to hundreds of terabytes, he says.

Copan Systems, another startup to come out of the closet recently, also emphasizes massive capacities: A fully configured Revolution 200T can house 224TB in a single cabinet. Targeted at backup and scientific applications, Copan advertises its disk-based system as cost-competitive wi...

For Hewlett-Packard, which entered the fixed content fray last month with its Reference Information Storage System (RISS), scalability is twofold: It has to scale in terms of capacity, of course, but just as important, it must scale its performance. RISS, which is built around "storage grid" technology acquired last year from Persist Technologies, scales thanks to individual "smart cells" that make up the system, each containing processing, network connectivity and storage capacity.

Ultimately, says Calvin Zito, HP's information lifecycle management (ILM) strategic marketing manager, archive storage isn't about building a "digital landfill" out of cheap disk--it's about "how do you access the data?"

With RISS, Zito explains, accessing information is speedy, thanks to the platform's high-performance indexing and search capabilities. Furthermore, RISS "harvests" application and stores it in an industry standard format (e.g., XML) to ensure that the data can still be accessed "ten years from now, when the application used to create it is gone."