Snapshots and backups. You hear these terms thrown around a lot, and sometimes interchangeably. But while they may sound similar, there are some important differences. Both can provide a way to recover systems and data to their original state at a point in time, but they go about it in different ways.
What is a Backup?
If you're reading this blog, it's safe to assume you know what backup means. As a noun, essentially, it’s a copy of your data and files. Traditionally, that copy is placed in a different location than the original content -- making it ideal for disaster recovery or to restore servers after a server failure, or more commonly restore data accidentally deleted by users or in some cases maliciously corrupted by ransomware. As a verb, backup is the process of creating and maintaining multiple backup copies of your data. Depending on the type and amount of data, backups can take minutes, hours or even days to complete, meaning there may end up being significant differences between your last complete backup, based on when the backup started and your live production data.
For this reason, continuous replication of data is often desirable to reduce data loss in the event of a disaster to meet corporate recovery point objectives measured in minutes vs. more typical once a night backup schedules. Often, backups are stored for long periods of time and may serve the function of an archive for various purposes such as legal discovery and compliance. In many cases, cloud based storage can be a cost effective approach for long term backup data retention however, it is important to consider the time that would be required to access large amounts of cloud based data for recovery Cloud is generally not suitable for recent backup copies that are most likely to be needed to recover systems and data quickly to meet corporate recovery time objectives (RTO.)
What is a Snapshot?
As the name implies, snapshots provide a quick “picture” of a server (including its files, software and settings) at a particular point in time. Generally snapshots are instant, and preserve a point in time state without having to move or copy existing data at all. For this reason - most modern backup approaches actually rely on snapshot technology to provide a stable, unchanging point in time image to do a backup of vs. doing backup of the live system and data and having to contend with open files and active applications changing data during the backup. In systems such as HC3, snapshots can also be replicated to a secondary system located in another location or even a cloud based service (DRaaS).
When using a snapshot to restore a server, that server will revert back to the exact state it was in when the snapshot was taken. For that reason, snapshots are a quick and effective way to “roll back” to that point in time. Because snapshots don’t make a full copy of all data, they can be very space efficient, and in the case of HC3 a snapshot actually consumes nearly zero additional storage at the time it is taken. However as data blocks in the live system change through normal production use, the snapshot needs to preserve the original snapshot content as well which means that the storage space attributed to a snapshot will grow over time as the snapshot data diverges from the live system. Therefore, snapshots are generally best for shorter term storage retention. (for more detail on HC3 full VM snapshots see the HC3 Theory of Operations - link)
Snapshots at the VM level allow for easy, instant, point-in-time recovery, providing protection against ransomware, malware, and accidental file deletion. Snapshots can be used to recover entire VMs or individual files. Support for hundreds of snapshots per VM plus, flexible scheduling and retention options enable multiple levels of protection for VMs, based on the need for each individual VM.
Snapshot vs. Backup - Differences and Similarities
So, to be clear, snapshots alone are not true, independent backup copies, but they are very useful for certain purposes, like development and testing. For instance, a snapshot can be used as a quick failsafe in case you need to rollback a patch, an upgrade, a test, or any operation that might be deemed “unsafe.” Snapshots are also a great way to recover from cybersecurity attacks. With the right type of virtualized systems such as HC3, infected virtual machines can be reverted back to pre-ransomware snapshots and be up and running within minutes.
Destination storage flexibility is also part of the differentiation. Snapshots use the same storage stack as the primary runtime environment. Often, the primary system storage stack is very good, very fast, and very expensive by comparison to off platform storage options. That is part of what makes Snapshots not terribly practical for some backup use, especially for longer term retention.
Backups, on the other hand, can take advantage of multiple storage media types ranging from inexpensive NAS, tape, and cloud. In doing so, restoring from backups loses some of the immediacy of restore in favor of flexibility of placement and retention, which allows for much longer term archival storage schemes. Additionally, depending on the media type used as a target, backups allow for the air-gapping of the backed up data away from runtime events. This acts as an additional layer of protection and isolation from some malware threats in today's world.
In fact, Scale Computing HC3 provides the ability to protect every virtual workload with snapshots to quickly recover in case of an attack. If you’re hit by ransomware, you can simply revert to a previous snapshot and carry on. Of course, hyperconverged systems like HC3 are already well-suited to withstand attacks simply because they more tightly integrate servers, storage, virtualization and data protection features than traditional systems that combine these components from different vendors using open network protocols. This tight integration reduces the attack surface of the system. So, hyperconvergence is an innovative way to simplify your IT operations and is more effective at protecting data than traditional systems.
The HC3 Data Protection Suite built into HC3 provides native backup and disaster recovery capabilities including replication, failover and failback and flexible snapshot scheduling. As highly available clusters or more cost-effective single appliances, HC3 systems make ideal backup systems for your offsite disaster recovery. And for users whose requirements go beyond that, we provide Acronis Backup Advanced, expanding backup and recovery options as well as active malware and ransomware protection within and beyond their HC3 system.
We also recently added support for Acronis SCS backup and recovery solutions on HC3, which provides a higher level of protection for U.S. public sector organizations including federal, state and local government, public utilities, and education, healthcare, and nonprofit organizations.
Snapshots and backups are different, but they both serve a purpose. And they’re not an either/or proposition -- there is nothing preventing you from using both, and you should.