Table storage is built to handle web scale data, allowing intensive and fast read and write transactions. A single account has a target throughput of 20,000 requests per second, while a single table partition can handle up to 2,000 transactions per second.

The service offers a cost-effective storage solution where you only pay for what you use and at a lower cost than on-premises storage options. As of writing this book, the average cost of 1 TB in a table with a local redundancy option is 6.5 cents (US), and 8 cents (US) for a geo-redundant one. Also, Table storage is exposed through a REST API that makes it always available to any client that can make and handle HTTP requests. You can build your own libraries to manage connecting to the REST service, or most probably, you will use one of the open source libraries that are available for most of the development stacks.

Tables are non-relational, which means you cannot have relations and foreign...