As all of us know, an application comprises broadly of 3 tiers, the presentation, the business tier and the database. The Web 2.0 phenomenon has had a huge impact on the way applications are designed and developed. The most obvious changes are on the presentation layer because that’s what’s visible, but the business tier has also seen major changes, mainly in terms of exposing data as API, RSS feeds etc in order to open-up the closed data and making it ready for reuse.

However, what hasn’t changed so much is the database layer. All the new Web 2.0 features are still built on top of existing relation database, be it Oracle, MySql, Sybase or any other similar product. In fact, even before Web 2.0 phenomenon kicked in, there were huge changes to the way applications are written (use of object oriented languages for example), but database layer has pretty much remained oblivious to all these developments over the last couple of decades. To be precise, companies like Oracle have been working hard to introduce some features that make database programming closer to object oriented paradigm, but I think most would agree that fundamentally RDBMS (Relational Database Management System) hasn’t changed a lot. The object oriented database approach tried to address this, but it wasn’t popular mainly because of mixing objects (which belong to business tier) with data layer, rather immature object query languages and performance issues.

RDBMS offers a lot of rich functionality which are not easy to replace with anything that’s available today. However, there are situations when the complex nature of RDBMS can actually lead to more cost especially if someone does not want to use all the advanced feature it offers. Amazon’s recent launch of SimpleDB is targeting such users and trying to fundamentally change the way data is managed and requested using Database-as-service paradigm.

The SimpleDB let’s you store data in “domains”. The domains can store information about an item as attribute than can have multiple values, so that does not exactly fit into thinking in terms of row/column that we are used to. Also, it let’s you store different set of attributes for different items, which kind of translates to having different number of columns for different rows in the conventional tables - it sure is going to take a while before people get used to it.

Amazon indexes the data behind the scene to make the search faster. The querying supports parameters like =, !=, <, > <=, >=, STARTS-WITH, AND, OR, NOT, INTERSECTION AND UNION.

In its current state, I don’t think any big enterprise will seriously consider this option but it can be quite lucrative for small businesses and startups. It takes away all the pains of acquiring hardware, hosting clustered RDBMS, hiring database administrators and high cost associated with all this. Amazon is offering data storage at $1.50 per GB per month with additional costs for bandwidth and CPU usage, but overall it’s definitely much cheaper than conventional options and it let’s companies to scale up as and when they need it! SimpleDB combined with Amazon S3 (for storing large chunks of un-indexed data) offers a very cost effective and scalable data storage solutions for small firms. This product seems very well planned and integrated with their other Software(and Hardware)-as-service offerings.

As the name suggests, it’s “simple” but I feel it still can have far reaching effect on how people think about managing and accessing data. This will also get many brains thinking about alternatives to conventional RDBMS and I think that’s great news!

What’s your take on this alternative to conventional databases?