Amazon, the same company that sells books and things, also sell all sorts of electronic internet services. Very basically they built their own hardware and software to sell books, realised that what they made others would like to use, that they had a lot of it and so could offer this cheaply, and so sold it. I guess it is a bit like electricity generation – Amazon need to be able to make every sale at Christmas time, so need a lot of capacity, but what to do with this the rest of the time? Offer it for sale, scale it up, and then let anyone who uses this (if they wish) use an ‘elastic’ service that automatically grows to cope with surges (e.g. a sale) and then goes back to normal the rest of the time.
These services (which get generically known as Amazon Web Services), run a lot of the web services that we use everyday, though we don’t know it. They are, pretty much, the infrastructure back end to web companies. A new service has just been announced, relevant to archives the world over. It is called Amazon Glacier and it is “reliable and extremely low cost storage service designed for data archiving and backup”. Here the model is not lots of access just parking it somewhere safely. Now they charge a monthly fee per gigabyte for storing it and then additional fees if you download. To store is 0.01 a Gig a month, which is in the scheme of things a crazy prize (if you can remember when a 640Kb floppy drive cost more than a dollar). For something like the Circus Oz archive where there is about 16 Terabytes of video (at very high archival quality resolution) this is a bargain, and would cost around $300 a month. If this was being downloaded it costs about 0.12c per GB, so one show a month (let’s say 6 GB) would be a few cents.
For a project such as this, just in terms of figuring out where to store the data, this is a big shift. The previous model that Amazon offered just for storage was in the order of $1600 per month. Oh, I guess it is called glacier since it is for big, slow things, that don’t move much.