Recovery is really the only reason why companies are investing in backup solutions, but meeting the recovery objectives of the business is becoming more challenging by the day. 39% of European IT managers confirmed in a recent IDC survey that their recovery objectives have decreased from hours to minutes over the past year. This is not only true for traditional business-critical applications such as ERP systems and databases - it now also includes collaboration and productivity tools for employees, as communication and interaction with co-workers, customer, and partners are done through multiple digital channels which need to be available at all times.
Mainstream adoption of virtualisation has made cloud services more popular for backup and recovery because virtual machines can be replicated fairly easily to the cloud provider site, especially when compression and deduplication are switched on.
When you add cloud to your backup and recover process, which is currently a priority for 36% of European IT managers according to IDC research, there are a few points that you need to consider to ensure effective recovery:
- Do you have complete flexibility with multiple recovery options? Recovery cost and speed depend on the size of the data volume to recover and the available network bandwidth. When you want to recover from the cloud, it is important that you have granular options of what exactly you want to recover to keep recovery cost down and ensure fast recovery. If you only want to recover a single file, it does not make sense if you have to recover an entire application or, even worse, a full server to access it because you will be charged for the entire data volume that you are recovering. While file recovery over the network link is relatively speedy, recovery of an application or a full server takes many hours or sometimes even days, depending on its size. This is also true when you want to recover files or applications that run in virtual machines. File recovery is a particularly good candidate for self-service, so that employees can recover the files that they need easily and instantly through a self-service portal without involvement or IT and without driving up recover cost.
- Can you choose a hybrid recovery process? A common best practice is to have a hybrid architecture where the latest two weeks of backup are stored onsite and are replicated to the cloud service asynchronously. Only long-term storage of backups takes place in the cloud, as cloud services are typically optimised for capacity and low cost, and not for performance. So to ensure fast recovery in the case of a full system failure, organisations can recover from their performance-optimised on-premises backup system, while recovery of older files or data can happen from the cloud, which has comparable price/performance characteristics to a tape infrastructure.
- Can you reduce data transfer costs by using deduplication and compression? The network is the bottleneck in a recovery situation. Unlike the backup process, which is a trickle process where only small chunks of changed data get transferred to the cloud target every day, the recovery process is a burst process, where a large volume of data needs to be transferred at once. Effective data deduplication and compression are necessary to speed up data transfer over the network link and, most importantly, to keep the data volume that needs to be sent to and from the cloud platform as small as possible to keep the cost under control.
Overall you need to assess where cloud makes sense in your backup and recovery process and for which workloads you want to use it. If you decide to use cloud services as backup targets, you need to make sure that your backup solution optimises recovery from the cloud and helps you keep recovery costs under control. Using granular recovery options in combination with snapshots, data deduplication, and compression to shrink the data footprint that needs to be transferred over the network, you can ensure that only what is needed is recovered. Otherwise you might get a big surprise in the form of a higher than expected cloud service charge.