Researchers at the Massachusetts Institute of Technology's Media Lab, a pioneering research facility in the digital technologies field, will tap a 1.4 petabyte IP storage array to help aid a three-year study into early-childhood development.
The study, which began last year under the direction of MIT Associate Professor Deb Roy, involves the collection and analysis of several terabytes per week of digital audio and video recordings of early childhood behavior with the aim of better understanding how children learn and how language is acquired. MIT estimates that the data will exceed a petabyte by the middle of 2008.
MIT officials say the project is on the cutting edge of large-scale data systems management.
"I think what we're seeing is an indication of what the future will be like, with close collaboration on a large-scale project of global significance between the academic and industrial worlds," said Frank Moss, director of the MIT's Media Lab and a former chief executive officer of Tivoli Systems, a systems and network management firm acquired by IBM in the mid-90s.
"As the former head of Tivoli systems, I was responsible for helping customers deploy large-scale computing environments, and the biggest problem in making large-scale systems work is how they're managed. This project will provide unheard capabilities whereby petabyte-levels of storage can be managed by a few graduate students," Moss said.
Four technology companies have teamed up to construct the storage infrastructure to support the project. Zetera will provide its Z-SAN SOIP (storage over IP) architecture as the core technology. The infrastructure will also include more than 300 Bell Microproducts Z-Rack storage enclosures, which are built on Zetera's Z-SAN technology; 100 10-gigabit Ethernet switches and 400 blade processors from Marvell Technology Group; and about 3,000 SATA (Serial Advanced Technology Attachment) hard disk drives from Seagate Technology.
MIT says speech and video mining technologies used in the project could also provide applications for the enterprise computing world in such industries as multimedia data management, business intelligence and security.