Robin Good's insight:
AcademicTorrents is a new web service which allows any organization owning large datasets (no size limits) to easily distribute them without needing a dedicated infrastructure.
The brainchild of Joseph Cohen and Henry Lo, two PhD students working at the University of Massachusetts in Boston, Academic Torrents facilitates the job of researchers, journalists and information analysts in finding, accessing, curating and downloading large-size datasets.
Technically-speaking AcademicTorrents is a bittorrent-type redundant high-speed network and a full distributed system for sharing enormous datasets. As a P2P system it doesn't require intermediate servers, is also fully scalable, secure, fault-tolerant and can act as a reliable repository for data allowing fast downloads.
Users can also search the full index, and can create curated datasets collections containing any kind of files and which can be downloaded as a full bundle.
This type of system could prove to be an excellent resource for libraries storing digital papers as they would store books, and for simplifying the distribution requirements of any organization needing to publish, curate and share large datasets.
"A robust distributed replication design allows libraries to utilize this system as their backbone. Providing fault tolerant hosting of curated data for a university, research lab, or home library.
...Also, this system can be used as the foundation of a new open-access publishing system where libraries manage data instead of licenses for external data sources."
Find out more: http://academictorrents.com/
More info: http://academictorrents.com/about.php
Browse Datasets: http://academictorrents.com/browse.php?cat=6
Browse Papers: http://academictorrents.com/browse.php?cat=5