Data center protection is a mature discipline, but still one of the most challenging tasks facing data centers today. They are constantly under pressure to deliver faster recoveries. Though the market is mature, there is a new wave of hardware and software options vendors are offering to help IT professionals better meet their data protection and recovery challenges.
When disk appliances first appeared on the scene more than 15 years ago, they were used as a gateway or cache to the primary backup target, often a tape library. In the early 2000s, disk appliances added deduplication and compression, making disk storage almost as affordable as tape. As a result, most data centers view a disk backup appliance as the primary and, in many cases, only backup storage device.
Recently vendors have added more intelligence to their backup appliances to differentiate them from competitors and help customers improve the data center protection process. Vendors such as EMC/Data Domain and ExaGrid have added application-specific integration so they can better handle backups from mission critical environments like Oracle and MS-SQL. Some vendors have even added the capability to host components of backup applications on their systems to reduce network traffic and improve performance.
New vendors like Cohesity and Rubrik have also appeared on the market, leveraging a distributed computing model similar to Amazon and Google but applying it to secondary storage. The result is that these storage systems not only scale, but are also designed with data protection in mind. This new breed of secondary storage systems converges security software -- which can be used for data center protection -- similarly to how hyper-converged storage converges compute, networking and storage.
Copy data management
Another challenge facing data centers is all the secondary copies of data that are created to support functions like test, development, reporting, analytics and data center protection. Not only does this secondary set of copies consume storage -- as much as 10 times the primary copy -- it also consumes IT administration time to create and position the copy to the right process. Copy data management tools are designed to help IT professionals better manage extraneous copies.
Copy data management works by creating a single secondary copy of production data. When a particular application needs a copy of that data, it is given a snapshot instance from which to work. In the case of backup, the snapshot can be read-only, and the backup application can copy that data to the appropriate disk or tape backup resource. In the case of testing or development, a writeable snapshot can be created and assigned to those tasks.
Copy data management should also automate the creation of the snapshot data and its presentation to the requesting process. For example, if test or development wants a copy of the production data every four hours, the copy data management tool can be programmed to snapshot the current backup copy for version retention, update the primary copy with the latest changes from production, then create a snapshot of the updated copy and move it to test or development. The test and development teams should be able to refresh their environment and access this new copy.
Recovery-in-place and replication
Users are demanding faster recovery times regardless of what causes an outage. A key technology to address these demands is recovery-in-place. Available from vendors such as Veeam and Nakivo, recovery-in-place allows a virtual machine's data store to be accessed directly from the backup storage device. Recovering on the backup storage device eliminates the time required to move the system across the network.
To further reduce recovery time, many backup application vendors are beginning to include replication. Replication is used both locally and remotely to protect the most critical applications. Unlike recovery-in-place, replication captures production data more frequently and stores it in a native state in a secondary array. While it does require more disk space, the data is essentially live and ready for access. The virtual machine just needs to point at this copy of data, and after a quick data integrity check, it should be ready to resume operations.
Thanks to virtualization, excessive data growth and increased recovery pressures from end-users, data center protection is a more difficult task than ever. The good news is that new technologies are available that can meet data growth demands as well as simplify this increasingly complex process. The key is to combine hardware and software so that more of the process is centralized and automated.
How to make recovery-in-place work for you
Data protection issues soothed by copy data management
Defining copy data management as a data protection approach