From 3-2-1 to versioning
- 3-2-1 rule
- Keep three copies of your data, on two different kinds of storage, with one copy off site. Today that off-site copy should also be locked or offline.
Where this shows up in our work - Agent
- A small program on each laptop, desktop or server that runs its backups on schedule and reports back. If an agent stops, the device stops backing up, so someone has to notice the silence.
Where this shows up in our work - Air gap
- A copy with no network connection to your live systems, such as a drive kept unplugged in a safe. A locked cloud copy that nobody can change or delete serves the same purpose without the drive.
Where this shows up in our work - Application-consistent backup
- A backup taken after a database or app has been told to finish writing, so the copy opens cleanly. Without it, a restored accounting database can come back half-written.
Where this shows up in our work - Bare-metal restore
- Rebuilding a whole machine from its image backup onto new or wiped hardware, with no operating system installed first. It is how a server with a dead disk comes back in one step.
Where this shows up in our work - Deduplication
- Storing each repeated block of data once, however many files or versions contain it. It keeps versioned backups affordable, and we bill extra server storage after it is applied.
Where this shows up in our work - Differential backup
- A copy of everything that changed since the last full backup. Each one grows until the next full backup, but a restore needs only two pieces: the full and the latest differential.
- Encryption at rest
- Data scrambled while it sits in storage, so a stolen disk or a leaked file is unreadable without the key. We use AES-256 for every stored copy.
Where this shows up in our work - Encryption in transit
- Data scrambled while it travels between your devices and storage, so nobody on the network path can read it. We use TLS for every transfer.
Where this shows up in our work - Encryption key
- The secret that turns encrypted data back into readable data. On Server Backup and Immutable Offsite Copy you can hold it yourself, but if you lose it, nobody can restore your data.
Where this shows up in our work - Failback
- Moving work from a temporary system back to the repaired or replaced original. It needs its own plan, because changes made during the outage have to come back too.
Where this shows up in our work - Failover
- Switching to a standby system when the main one fails, such as starting a failed server as a cloud virtual machine. Staff keep working while the hardware is fixed.
Where this shows up in our work - File backup
- A copy of documents and folders rather than the whole machine. It is quick for everyday restores, but it cannot rebuild an operating system or its apps.
Where this shows up in our work - Full backup
- A complete copy of everything selected, made from scratch. It is the starting point that incremental and differential backups build on.
- Image backup
- A copy of a whole disk: the operating system, apps, settings, permissions and data. It lets a failed machine be rebuilt in one step instead of reinstalled by hand.
Where this shows up in our work - Immutable backup
- A copy that nobody can change or delete until its lock period ends, administrators included. It protects backups from ransomware and from mistakes made with an admin password.
Where this shows up in our work - Incremental backup
- A copy of only what changed since the last backup of any kind. It is small and fast, which is how a laptop can back up every hour.
- NAS (network-attached storage)
- A storage box on the office network that shares folders with everyone. It makes a fast local copy, but it is still one copy in one building.
Where this shows up in our work - Off-site copy
- A backup kept away from the building where the live data lives. It should be far enough that one fire, flood or storm cannot reach both.
Where this shows up in our work - Ransomware
- Malicious software that encrypts files and demands payment for the key. CISA warns that many variants look for backups to delete or encrypt, which is why one copy should be locked or offline.
Where this shows up in our work - Recovery point objective (RPO)
- How much work you could lose, measured as the time back to the last good copy. An hourly backup means a recovery point of about an hour.
Where this shows up in our work - Recovery time objective (RTO)
- How long you can wait for a system to be working again. It decides whether a restore can run overnight or needs a standby system.
Where this shows up in our work - Recycle bin
- Where Microsoft 365, Google Workspace and computers hold deleted items for a limited time. Microsoft’s documentation says SharePoint keeps deleted items for 93 days; after that, only a separate backup can bring them back.
Where this shows up in our work - Restore drill
- A scheduled test in which real data is restored to an isolated environment, timed and written up. Ours restores a sample of files, one mailbox and one full system every quarter.
Where this shows up in our work - Retention
- How long old versions are kept before they are deleted. Our server default is 30 daily, 12 monthly and 7 yearly versions.
Where this shows up in our work - Runbook
- Written, step-by-step restore instructions for each system, in the order they should come back, with who to contact. It lets someone other than the usual IT person run a recovery.
Where this shows up in our work - Seeding
- Sending the first full copy of a large data set, sometimes on an encrypted drive shipped to the storage region, so later backups send only changes. It matters for offices with slow uploads.
Where this shows up in our work - Snapshot
- A point-in-time view of a disk or volume, usually kept on the same system. It is quick to roll back to, but it fails along with that system, so it is not a backup on its own.
- Sync
- Keeping the same files on several devices, so a change in one place appears everywhere. Deletions and encrypted files spread too, which is why sync is not backup.
Where this shows up in our work - Tabletop exercise
- A recovery rehearsed around a table, talking through each step of the runbook without touching live systems. It finds missing steps and missing contacts before a real outage does.
Where this shows up in our work - Versioning
- Keeping earlier saved versions of a file, so you can go back to how it looked last Tuesday. Team Sync keeps 180 days of versions.
Where this shows up in our work
Put these terms to work
A free 45-minute backup check shows what you copy today, runs one test restore, and puts the findings on one written page.