Azure table storage is an extremely useful feature of the Microsoft Windows Azure offering. The cost per transaction is extremely low compared to other providers and the cost per amount of storage used is also very appealing.

Azure table storage is contained within an Azure storage account, along with the other storage entities: blobs and queues. You may have multiple Azure storage accounts set up within your Azure account, each with their own keys and unique set of URLs for the contained blobs, tables and queues. An Azure storage account is a discrete Internet based storage solution that can be used by your Azure applications, other non-Azure websites or stand-alone applications within your local environment. Your application simply needs either Azure connection strings containing the appropriate access keys or needs to be accessing parts of the storage account that have been given public visibility.

Whilst Azure table storage may not be the most sophisticated storage system ever devised, it is perfectly capable of the more basic operations involved in a distributed system. For example, see our Logging Using Azure Storage Tables article for information on how Azure table storage can be leveraged for logging purposes. Azure table storage is never going to be a substitute for a large scale relational database system like Azure SQL Database (previously SQL Azure). However, if your requirements do not need a full relational database, the cost advantage of table storage is overwhelming. Azure table storage is perfectly suited to linear queries, queries requiring the whole data set to be returned or simple atomic actions to suffix data onto an existing data set. At present there is no ability to add secondary indexing and the only index you can use is the row key (which is in effect the primary key for the table).

Therefore, if your requirement is simple data capture or a system that needs to read the entire result set of a table, Azure table storage could be a cost effective solution for you. For example, we use Azure table storage here for managing our CMS (Content Management System) of articles. Because we need a complete index of the articles available, the full table is read, retrieving the columns that are important for our index pages (e.g. title, date/time, etc.). When someone selects an article to read, that record is retrieved from its row key and displayed.

For some real life code examples of using Azure table storage, have a look at our using Azure table storage for logging article.