Backup Considerations for Virtualized Environments
You cannot just rely on snapshots of your virtual machines even if these are taken on regular basis without having a proper data backup plan in action. Your data backup strategy should be based on a reliable and fast backup/restore solution which at least caters for the business most critical data. What do I mean with a reliable backup/restore solution? - A data backup solution which integrates seamlessly with the underlying hypervisor and one which is regarded trustworthy by fellow system administrators. What do I mean by a fast backup/restore solution? - A data backup solution that does not compromise hypervisor performance and runs autonomously as possible. Some products provide agent less solutions and can backup and restore multiple VMs simultaneously.
The following checklist of features would give you a head start when looking for a backup solution for your virtualized environment:
Hypervisor Management Integration – the integration with the hypervisor management console is a big plus as it helps you relate the various components better, for example, a simple right click option from within the same management console that you use for the daily tasks might include features such as, job status, history, scheduling or even start/stop a backup job.
Ease of use – I have used some backup solutions for years and never got a full understanding of some of the features they had, and also, I came across some vendors that use strange terminology to refer to basic backup terms. It is imperative to feel confident that such important solutions are functioning as they should be hence, the design and ease of use helps in eliminating uncertainties. For example, a popular approach is to have a backup solution that runs as a virtual backup appliance. Although, this may impact on performance but it may give greater flexibility in most environments.
Technology used – the trend nowadays is more and more based on de-duplication technology. Traditional compression is still valid and plays an important role in saving storage requirements; however, the latest data de-deplucation methods are providing impressive achievements while most implementations improve the overall processing speed.
Verification features – the most worrying aspect of backup solutions is when we don't feel confident that the backed up data is complete that is, a solution needs to perform data integrity checks during both backup and restore processes.
Constraints – if an organization has a backup strategy in place then check that the backup solution can meet both RPO (Recovery Point Objective) and RTO (Retention Time Objective) requirements. A backup solution which is less flexible may present more constraints. However, remember that a good solution should provide support for various storage options and locations such as, tape drives and NFS connections.
Granularity – the extent to which an individual item/object or file can be backed up or recovered. It is useless to have the fastest backup solution that just backs up a whole VM as an image. If your environment dictates the restore of individual files then you need this functionality as well. A quick backup job may require the selection of a group of files, therefore File level backup/restore is an important feature.
There are more features that you would consider when procuring a backup solution such as, integration with Disaster Recovery and archiving plans, etc. However, the above points are key to most common environments. Finally, never forget to find a solution that has adequate monitoring and notification features that would alert you when something goes wrong.