Jesse Peterson

Amazon announced on Thursday (12/13/2007) this week a new service: Amazon SimpleDB. SimpleDB is a new part of the Amazon Web Service (AWS) offerings. This hosted application provides, as its name suggests, a simple database. It isn’t a relational database. It instead stores attributes-value pairs. The data values are indexed. SimpleDB has a special query language that lets you efficiently query the attribute values. SimpleDB joins the Amazon Elastic Compute Cloud (EC2) and Amazon Simple Storage Service (S3). Like these other services, SimpleDB is priced on how much you use it: so much to put items into SimpleDB, so much to read items out of SimpleDB, so much for the time to execute a query, etc.

One of the recommended practices of SimpleDB is to store the metadata in SimpleDB and the large data in S3. For instance, you could upload your FLAC encoded music library to S3. Each audio file has metadata, like an artist, album, title, track, genre, year, etc. This metadata could be entered as items within SimpleDB. Each item would also include a URL to the actual FLAC audio file stored in S3. You could then run queries against SimpleDB to find different music that matches your criteria. For instance, find all items where the “artist” is “Nerf Herder“. This would return all items, which would include the URL to the audio file, from your music collection that were performed by Nerf Herder. You could then have you audio player on your laptop play the audio, retrieving it from S3. This would allow you to access your music from anywhere and have it stored in a highly reliable system in case your CD was scratched.

I think that this could be a really neat system. Though, it does remind me more and more of the “old” mainframe systems. (Okay. I am really not that old. At Pearson where I currently work, they still use a mainframe.) I have been interested in S3 for some time, like my littles3 implementation, so I will probably keep my eye out for interesting uses of SimpleDB. I also wonder what Google will do to have something like this. I have heard rumors about Google’s “GDrive” for over a year now.

Now if I can only find some reason to justify using SimpleDB once it is available. (It is in limited beta right now.) 🙂

(Oh! My little example of “outsourcing” the storage of your music catalog from above could also be related to the book that I am currently reading, “The 4-Hour Workweek: Escape 9-5, Live Anywhere, and Join the New Rich“.)