Windows Azure Recovery Services. Part 1.

by Pavlo Revenkov

Oftentimes while working with the Windows Azure cloud platform, clients require a backup service for data stored either in the cloud or on the local servers. If we are looking at the SQL Azure database, then the answer is simple - SQL Azure Data Sync. But what do we do if this functionality is required, for example, for virtual machines? Or maybe for data that is not relational or is not stored in a relational database?

In April 2014, the Microsoft team developing the Windows Azure cloud platform announced a standalone service meant to handle this common issue. It is called Recovery Services. This service allows saving data backups from machines running under Windows Servers.

The point of this service is extremely simple. You install a special application (Windows Azure Backup Agent) on a machine that is running under a Windows Server and contains data that should be stored in the cloud. This application synchronizes the selected data according to a schedule (can be run manually). You can create backups of entire partitions, as well as separate directories and files.


As you know, a large amount of data, let's say backups, can be put in a storage service provided by Windows Azure. However, to access it, you must use either Windows Azure SDK (if we are developing our own solution), or third-party solutions (for example, Azure Storage Backup). I would much prefer to have this functionality provided out of the box.

Let's compare the capabilities of Windows Azure Recovery Services and Windows Azure Storage:

* - Geo-replication. The ability to replicate data within the confines of a region. Each region, for example Europe, contains two data centers (sub-regions): West Europe and North Europe. Your data is replicated in the sub-regions; in other words, the data located in the West Europe region storage is replicated in the North Europe region, and vice versa.

As you can see from the table, the Windows Azure Storage services are available in a larger number of data centers; they are also cheaper and support geo-replication. However, Recovery Services have none of these benefits. On the other hand, work with Recovery Services is not rated via transactions. Also, keep in mind that the entire Recovery Services setup process boils down to installing additional software on a machine with a Windows Server, while implementing the Windows Azure Storage data backup functionality requires using third-party software or developing your own.

Let's put together a simple table and compare the costs of storing backups in Windows Azure Storage and using Recovery Services.

* - salary level according to a study conducted for Kiev. Median value; experience - 1-2 years.

** - seeing as Windows Azure Recovery Services are currently in the Preview stage, the price has been reduced by 50%.

After my article dedicated to the SQL Reporting services had been published I was accused of not taking into account the potential costs of hiring a specialist to maintain the corresponding technologies. This time I tried to keep this point in mind.

As you can see from the table, the "net" cost of Storage Services is significantly lower than Recovery Services, and the larger the amount of stored data, the bigger the difference. However, if you take into account the salaries of maintenance specialists supporting each solution, Recovery Services becomes the cheaper option. But in this case, as the volume of stored data grows, the difference also shrinks.

The logical conclusion would be that the service should be chosen according to the needs of each particular task.

Next time, we will observe the process of setting up Recovery Services for Windows Servers. Thank you!