Photo credit Flickr user Tom Hynds, https://flic.kr/p/4NofNh
Do you speak tech? The tech world is full of so many industry specific terms, it’s hard to be fluent. Not to mention all the different dialects, with terms specific to different products or solutions. After a while, it can all sound like a bunch of buzzwords and jargon.
The backup industry is guilty of this too. The terms used to describe different types of backups, different methods of performing these backups and other very niche specific terms get thrown around all the time in both marketing and technical materials. In order to make it easier to understand what the heck backup vendors are talking about, we’ve put together this glossary of terms. Think of it as your backup dictionary with terms explained in plain English.
Full volume backup – Just like it sounds, a full volume backup is a backup of an entire volume. These backups are generally large, but are also usually easy to restore from, and offer more restore options when compared to file-based backups. These backups also almost always have a local copy of the backup along with an offsite component, allowing for quicker restore times.
File-based backup – A file-based backup will backup files and folders only. This means that the backup is generally more lightweight, but also only backs up data files, meaning installations of both operating systems and programs need to be redone manually, and then the data put back into the software. These backups often don’t have a local copy, meaning any restores need to be downloaded from the cloud. The amount of time this takes can vary depending on how fast the download speed of the computer is and how much data needs to be restored.
Differential backup – A differential backup looks at the changes between the last full backup and writes those changes as a differential file. The advantage to these types of backups is that they’re only one step away from a full backup, meaning they can be faster and more stable. However, each backup is bigger than the last, and you need to make a new full backup with some regular frequency.
Incremental backup – Similar to differential backups, but instead of the difference between the last full backup and now, it’s the difference between the last backup of any kind and now. This is beneficial because large data shifts are only backed up once, unlike with differential where these large shifts would be copied multiple times in each backup until the next full backup includes it.
Continuous incrementals – Incremental backups commonly will create a new full backup every week or every month, which is very manageable when data sets are small and there’s no offsite in play. However, with larger data sets or when also offsiting backups, a new full backup every week or month quickly becomes impractical. To get around this problem, some backup solutions use a chain of continuous incrementals paired with consolidation of backup images or using reverse incrementals.
Intra-daily/daily/weekly/monthly/rolling files and consolidation – Rather than leaving a continuous incremental chain that grows into the thousands, most backup solutions offer consolidation of the incremental files in a backup chain, creating consolidated daily, weekly, and monthly files. These act as a kind of shortcut back to the base image, meaning even months out, there’s only eight or so image files used to mount the most recent backup.
Reverse incrementals – Instead of keeping a static base image and having endless continuous incrementals, some backup solutions instead inject the most recent incremental file into the full backup, while still keeping a chain of incrementals that work backward through time. This is advantageous because it means the most recent restore point is always a single file, resulting in better speeds when virtually mounting the backup or restoring from the most recent backups.
Data de-duplication – Data de-duplication is a method used to reduce backup size by flagging data that already exists and skipping over/removing that data from the current backup, because it already exists. While this is great for creating smaller backups, meaning less storage used and less data to replicate offsite, there are complications with de-duplication and encryption, because data de-duplication tries to find patterns in data while encryption attempts to hide any patterns in data.
Backup and Disaster Recovery Machine (BDR) – A frequently used acronym, a BDR is a server that acts as the primary storage for backups, a central point of control for managing the backups locally and also has enough power to virtualize backups temporarily if needed.
Virtualizing backups – Many full image backup solutions offer the ability to virtualize the backup image files directly, allowing for very fast recovery time objectives, or RTOs. While this is great for a short term recovery, using backup images as a permanent virtualized solution is unwise due to the multi-file nature of the solution. Most FIB solutions also offer permanent virtualization by allowing the user to convert a backup chain into a single virtual hard drive file to be used inside of a virtual machine.
Virtualizing offsite backups and VREs – There are many backup providers that offer offsite virtualization, and this can be accomplished in numerous ways. Gillware’s offering uses a Virtual Recovery Environment, a Windows machine running offsite that our partners can access 24/7 to virtualize their backups. Offsite virtualization can be useful during disaster scenarios, but can also have complications. For example, if the entire infrastructure for workers’ computers is gone, there’s often no computers left for them to use in conjunction with the offsite virtualization. Networking can also be a large hurdle to overcome.
Volume Shadow Copy Service – Most full image backup solutions on Windows machines use the Volume Snapshot Service to do their backups, also known as the Volume Snapshot Service, or VSS. VSS operates at the block level, and creates read-only copies of a volume, allowing for both consistent backups, and backups of files that are in by reading data from these copies rather than the live disk.
Virtual machine backups – Virtual machines can be backed up by both file-based backups and full-volume backups running on the VM itself, but also have a third option for backups done at the virtualization layer of type 1 hypervisors. These solutions are useful in cases where there are a large number of VMs on one machine, because a single install can back up a large number of mission-critical machines at the cost of only once license. The downside to these solutions is they often are billed per machine being backed up, so they can be costly when there are a number of VMs to back up.
Bare Metal Restore – The restore process for files and folders is usually very simple and done through a program running on a standard operating system. Full volume restores of the operating systems themselves, however, usually need to be done from a bootable OS created by the backup company, usually called a recovery environment. In these recovery environments, the option to write the backup images to a volume is provided, along with many other useful tools, like hardware independent restores and fixing boot issues.