Azure’s Table Storage is an interesting Beast. I set out to build a sample application with it, thinking it stored blobs of JSON or unstructured classes. It does support storing rows in a signle table where different rows contain different columns, but it’s not nearly as flexible as a NoSQL DB like MongoDB, which I thought was the case. A better comparison to Azure Table Storage is Amazon’s Simple DB.
The Structure is certainly simple to use, and highly scalable, but it’s not nearly as flexible as a Document-Oriented store like MongoDB.
I’m uncertain what are best practices for designing a Azure Table Storage schema and Data Access tier. Cost of the Cloud service, such as data size or per-insertion costs are factors to account for. Also what are the ideal target applications for something a technology like this? It seems similar in some regards to Hadoop, but the Map-Reduce, Flume and other tools are not there yet.
After playing around with Azure Table Storage for a short period of time, it’s definitely easy to use and something I’ll keep in mind in future projects.