A few thoughts on Amazon S3
Hi all! If you are still reading this thanks for sticking with us even though we have done very little in the past month. The first three weeks back at uni have been hectic for James and I. Hopefully things will start to calm down a bit now and then we can start working on Juvely again! Yay!
Over the last week when I have had some spare minutes I have been looking more into using Amazon S3 pratically as a storage system for an application, as apposed to using in the background as our backup solution. The application I have been playing about with is written in Ruby on Rails so I used the AWS::S3 gem to get going. To ease things even more I used the Attachment Fu plugin which makes things even simpler! About 10 lines of extra code (if that?) and attachments are uploaded to S3!
One of the things I was rather shocked about, as I don’t remember reading about it anywhere, is that Amazon S3 supports BitTorrent downloads. Nothing special is needed, just adding “?torrent” to the end of the query string! As soon as all browsers include a BitTorrent client (I think only Opera does by default at the moment) this could potentially reduce bandwidth costs drastically! If there are no users seeding Amazon S3 will act as a seeder so if there is support there is really no reason not to use this. So who is willing to write a script to detect whether the browser supports BitTorrent or not and passing the relavent S3 url?
Lastly, how would you like to save 15 cents? Although Amazon S3 is very cheap it still costs money during testing you may very well have to transfer huge amounts of data, or just be testing on a machine with a poor internet connection. Well no fear, meet Park Place! A clone of Amazon S3 written in Ruby! It doesn’t support everything under the sun and has a couple of bugs, but it seems fine for most uses!
Anyway hopefully this will help someone out who wants to experiment with Amazon S3! If you discover any useful tools any other hidden features add to the comments!