There are too many traditional backup products trying to tackle a nontraditional backup problem. As a result, they continue to use some of the old world backup techniques in the very new world of IT. These techniques waste storage space, network bandwidth and computing power. Let’s take a look at the two main ways they do that.
The Full Backup – A Holdover from Days Gone By
Whether you are backing up Hyper-V at the Hyper-V-level or at the VM level, you are most likely still performing backups that have a feature leftover from the tape era – the occasional full backup. Most backup software products were designed during the days when backups were sent to tape. In fact, many of these products were written during a time when very few customers used tape libraries. During the tape days it was very important to limit the number of tapes that were needed for an individual restore. This was even more true if a customer was not using automation, because every tape had to be manually put into the tape drive. I remember the days of literally sitting in front of a tape drive being forced to swap tape after tape. What a pain.
The best way to limit the number of tapes needed for a restore was to perform a regular full backup, typically once a week. Most people did this on the weekend when fewer people were using the servers, because a full backup impacted the performance of the server and the network. The weekly full backup became a standard which continues to this day.
In the last 10 years or so, backup to tape has become much less common, basically non-existent in the mid-market. Environments like Hyper-V are typically backed up to disk. So one would think that the idea of a full backup has been done away with. Yet vendors still build backup software packages around this idea of an occasional full backup. This is because they build software around the idea of first restoring a full backup and then restoring the associated incremental backups – like we did back in the day. If one never performed a full backup, the software would eventually not be able to perform a restore during a reasonable amount of time.
One idea that some backup software products have started using is a synthetic full backup, where a full backup is synthetically created from incremental backups. A backup system that has performed an incremental backup already knows all of the files that are still present and any files that are no longer there. This means it has all of the knowledge it needs in order to synthetically create a full backup from all of the incremental backups that it already has. Depending on its implementation, a synthetic full backup system can save storage space, computing power, and networking resources.
Full File Incremental Backups – Another Holdover
The second offender from a space-saving perspective is full-file incremental backups, defined as a backup that backs up an entire file if any part of that file changes. This has always been a space waster, but it’s a particular problem with applications like Hyper-V.
On one level, Hyper-V is just another Windows application. From a backup perspective, it’s a lot like a database application, because it uses a number of very large files (e.g. VDK files), all of which change every day. Just like a database application, when anything happens on a Hyper-V VM, it’s going to change the modification time and archive bit of the VDK files underneath that VM. This means that any application backing up a Hyper-V server without proper integration be performing the equivalent of a full backup every single time. This is why backup products must integrate with Hyper-V and perform block-level incremental backups.
Full backups must go, at least in the way we know them. They waste space, network bandwidth, and computing power. Synthetic full backups can be used without such negative impacts, if they’re designed in a way that minimizes these things. Ask your vendor about how their synthetic full backup saves backup storage space. If they say it doesn’t, then they only focused on the network and CPU savings. Look for a backup product that saves storage space as well.
Sponsored by NAKIVO
Headquartered in Silicon Valley, NAKIVO Inc. develops a fast, reliable, and affordable data protection solution for Hyper-V, VMware, and cloud environments. NAKIVO Backup & Replication provides native Hyper-V backup for Microsoft Hyper-V 2016 and 2012 R2. Out of the box, the product provides scheduled, image-based, application-aware, and forever-incremental Hyper-V backup. NAKIVO Backup & Replication stores backup data in the full synthetic mode. This enables to recover VMs from any recovery point even if a previous increment is lost or damaged. Over 10,000 companies are using NAKIVO Backup & Replication to protect and recover their data more efficiently and cost effectively. Visit www.nakivo.com to learn more.