The Microsoft Azure Partner Community is led by National Partner Technology Strategists from the Microsoft US Partner Team. Partner Community activities include blog posts, discussions on Yammer, newsletters, and community calls.
- Register for the October 16 US Azure Partner Technical Community call
- Read the US Azure Partner Community blog posts, including the September series about EMS
- Join the Azure Partners group on Yammer
- Sign up for the US Azure Partner newsletter
Microsoft Azure has changed the way we approach the age-old question about backup and recovery, but some of the fundamentals still apply. We talk a lot about Azure being backed up and redundant, but what does that really mean? In this post, I will talk about the Azure redundancy architecture and how Azure fits into a backup and recovery strategy.
Azure Data Redundancy
Notice that the title of this section is not “Azure Backup.” There is often a misconception that if a data file, a virtual machine, a service, or just about anything else is hosted on the Azure infrastructure, it is being backed up. While there are multiple copies of everything on Azure, they are not backups that can be restored.
In a traditional datacenter model, an organization uses a tool like System Center Data Protection Manager to back up its files and systems. A recovery time objective may be outlined, a retention policy defined, and then a backup schedule set. The environment will begin backing up based on the policies defined. In the event that a file changes and a user makes a request for a file to be restored, Administrators will use the Data Protection Manager console to locate the file and restore it. In this scenario, the file would be recovered and the user would be able to continue work.
Take a similar scenario with Azure in the mix. The same file mentioned above is hosted on Azure Blob Storage and Administrators don’t back up the data because it is hosted in Azure. When the user makes the request to restore the file, it is not able to be fulfilled. There is no Azure tool that allows Administrators to restore files, primarily because there is no backup to restore. When the file was changed, this data was replicated to ensure that the data would be highly available. Azure does not create or store a backup of the data.
Azure Redundancy Options
Azure provides high availability of data by using replication technology to create redundant copies of the data. The four current types of redundancy are outlined below and published on MSDN.
- Locally redundant storage (LRS) – provides storage redundancy in a single facility. Three copies of the data are replicated within a single datacenter. The risk with LRS is if this facility becomes unavailable, so does the data.
- Zone-redundant storage (ZRS) – addresses the risk of LRS by replicating the data three times across two separate facilities. These facilities may be in the same geographic region or across two separate region. This facility spanning is not selectable by the administrator.
- Geo-redundant storage (GRS) – is the default for storage account creation and again expands on the ZRS. GRS maintains six copies of the data spread across two geographic regions, three times in the primary region and 3 times in the secondary region. These regions are hundreds of miles apart ensuring that a single event cannot render the data unavailable.
- Read access geo-redundant storage (RA-GRS) is an expansion of GRS. With GRS the data is replicated across the regions but only one copy of the data is active and able to be accessed or modified at a time. RA-GRS allows data in the secondary region to be accessed.
On the Azure Storage Pricing web page, you’ll find an overview of the four storage options and their associated costs. There is a cost to the increased level or redundancy and access.
Azure Backup and Recovery Strategy
Azure can, and should, play an important role in a customer’s backup and recovery strategy. Providing and ensuring storage availability has evolved from backup to digital tapes, and then shipping those tapes offsite to backing up to a disk array to a remote datacenter. This evolution continues with organizations that back up to a disk array and then ship the backups to the cloud where they can take advantage of the decreasing cost of storage. Some organizations are even forgoing the primary disk array and backing up straight to the cloud. There are multiple ways to use Azure as part of an effective backup strategy.
Azure as a Backup Target
With the release of Windows Server 2008 R2 SP1, the ability to backup directly to Microsoft Azure was introduced and built right into the operating system. This option still exists today in Windows Server 2012 R2. SQL Server 2012 SP1 CU2 also has built-in options to backup directly to Azure. In SQL Server it is also known as SQL Server Backup to URL.
These backups use Azure Blob Storage as a target, but function as a traditional backup. Recovery from this type of backup would allow for point in time recovery. If we carry forward our use case from earlier in the post, a file would be able to be restored if a user requests to go back because of a change.
System Center Data Protection Manager
Azure backup, included within the products, provides a built-in solution but managing a schedule for individual servers can be complicated and heavily burden system administrators. System Center Data Protection Manager (DPM) provides a central backup tool for the enterprise, and has the ability to connect with Azure and leverage blob storage as a remote target.
DPM can be configured for the cloud in a few different Azure scenarios. Organizations that still leverage tape backups can add Azure to this and perform a Disk-to-Disk-to-tape (D2D2T) + Azure. For those not using tape, DPM can perform a Disk-to-Disk (D2D) + Azure backup, as in the diagram below. These solutions provide short-term, cloud-based storage location for backups. In connection with DPM, Azure can store up to 120 days of data when leveraging a daily synchronization.
For users leveraging DPM 2012 R2 with Update Rollup 3 (UR3), longer term retention can be achieved. With a monthly cloud synchronization, retention can be set up to 3,360 days. For more information about this announcement, check out the System Center: Data Protection Manager blog.
While most enterprises are looking to utilize the cloud heavily going forward, we understand there will still be the need for infrastructure in the datacenter. Part of this need will be for storage due to the explosive growth of data that organizations maintain. Microsoft StorSimple is a storage array that combines on-premises storage with the power of Azure, in a tiered storage solution.
The StorSimple device is installed in the organization’s datacenter with a minimal rack footprint, as little as 4U. The device contains SSD and HHD storage tiers inside the device. It is connected to Azure to provide an additional storage tier. The storage management is all handled by the device and presents to the servers as an iSCSI target. This means as data is added to the device, it will spill over to Azure automatically. The data in Azure is encrypted with a 256-bit AES key that only the organization has. In the event that a restore needs to be performed, the organization can install the key in another device or even a virtual device in Azure where the data can be restored.
The Partner Opportunity
While Azure can play a large part in a backup and recovery strategy, it is important to understand what is—and isn’t—built into the platform. Here are three key points to remember:
- Data redundancy is not a backup, but does ensure that data remains available. You can help your customers using Azure validate their backup strategy and design one should it be needed.
- Backup to Azure is becoming more and more integrated into Microsoft products but a centralized tool, like System Center Data Protection Manager can make backups to the cloud easier to administer. You can help customers use Azure as a backup location in numerous scenarios.
- StorSimple provides a great tool for backing up storage data directly to the cloud as well as providing a flexible storage solution for the datacenter. If you haven’t seen StorSimple, take some time to learn how it can meet your customer’s needs.