Organizations like Internet Archive and our treasured public libraries across the nation have increasing quantities of data to be stored. The LibriVox Audio Collection is a 22TB compilation of free public domain audiobooks that volunteers from around the world have recorded—over 16,000 public domain texts, including poetry, books, and short stories. Recently, Storj collaborated with Internet Archive to take advantage of the inherent benefits of the decentralized cloud to store the entire LibriVox Catalog.
Dominick Marino, a senior solutions architect at Storj, worked with Arkadiy Kukarkin, collaborations coordinator at Protocol Labs, to migrate the terabytes of files onto Storj DCS. Files uploaded to Storj DCS are automatically encrypted, split into 80 or more pieces, and distributed to more than 13,500 Storj nodes over the world. When someone wants to access the file, Storj DCS requires only 29 pieces to reconstitute the file—and they come from the nodes closest to the requestor. The closest 29 pieces to someone in San Francisco will be different from someone in Berlin—ensuring his the fastest file delivery at the edge.
Since there is no central point of failure, automated redundancy and 11 9s of durability, file loss is statistically unlikely (we have never lost a file!). And even if pieces were lost due to storage failure or nodes leaving the network, automated file audit and repair continuously ‘watches’ and quickly restores all 80 pieces of any file, guaranteeing our SLA of 99.95% availability.
Storj DCS not only costs pennies on the dollar, but overall TCO is significantly lower than today’s hyperscalers. Plus, listeners experience a faster download due to Storj DCS’s unique segment parallelism and edge delivery. Storj is excited to help safeguard the LibriVox Collection for years to come.
Watch the webinar to learn exactly how we did it.
