peshkova - Fotolia
Can you outline the different ways storage snapshots are created? Also, are snapshots a legitimate alternative to backup?
Storage snapshots can be created two different ways. The first method is called a differential snapshot. The basic idea behind this type of snapshot is that when a snapshot is created, the system actually creates a differencing disk. From that point on, all write operations are directed to the differencing disk rather than to the primary storage. This ensures that the primary storage remains unmodified so that the system can be reverted back to a previous state if necessary. Pointer-based snapshots are a newer type of snapshot, and they do not impact performance nearly as much as the differencing disk snapshots. The basic idea behind a pointer-based snapshot is that the snapshotting software assigns a pointer to each storage block. Whenever a file is modified, the storage blocks associated with that file are not overwritten. Instead, new storage blocks are created and the pointers are updated to point to the new storage blocks. If a rollback is required, then the software is able to determine which storage blocks were used at a particular point in time and then use those storage blocks to roll back the system.
First-generation storage snapshot solutions were never intended to be a backup replacement. They were more of a convenience feature than anything else. Modern snapshot solutions can be used as a backup replacement so long as they take measures to protect your data. It's helpful to remember that a snapshot is different from a data replica. With today's technology, it is nearly impossible to create a copy of a multiterabyte volume in a matter of a couple of seconds. Storage snapshots are not creating true copies of your data. Therefore, in order for a snapshotting solution to be a true alternative to a traditional backup, there needs to be a way of protecting both the snapshot and the original data. Typically, this means replicating the data and the snapshots to a mirrored location.
Dig Deeper on Disk-based backup
Related Q&A from Brien Posey
Hyper-converged architectures let users scale easily by adding nodes as needed, but it's important to be aware of vendor- and hypervisor-specific ...continue reading
Virtual desktop infrastructure environments are complex and have to work with many variables. A hyper-converged infrastructure can reduce costs and ...continue reading
It's possible to design hyper-converged hardware to run several hypervisors at the same time, but various factors make this a scenario we're unlikely...continue reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.