Data Backup and Recovery
RPO: how much can you lose?
Recovery Point Objective is the amount of work you are willing to redo. Nightly backups mean a failure at 4pm costs you a day of work. Hourly snapshots cost you an hour. The gap between those two is a business decision, not a technical one.
RTO: how long can you be down?
Recovery Time Objective is how long until people are working again. Restoring a file takes minutes. Rebuilding a failed server from scratch can take days. Knowing which you are set up for, before it happens, is the whole point of planning.
The rule that still holds
Three copies of your data, on two different types of media, with one copy offsite. It predates the cloud and survives it, because the failure modes it protects against have not changed.
3
Copies
The production data plus two backups. A single backup that fails during a restore leaves you with nothing.
2
Media types
Local and cloud, or disk and immutable storage. One failure mode should not be able to reach both copies.
1
Offsite
Geographically separate and not reachable with the same credentials, because ransomware looks for backups first.
What actually needs protecting
Servers and workstations
Full image backups, so a dead machine becomes a restore rather than a rebuild from memory.
Microsoft 365 data
Mail, OneDrive, SharePoint and Teams. Microsoft protects their infrastructure, not your deletions.
Line of business databases
Application aware backups, because a file level copy of a running database usually will not restore.
Configuration
Firewall rules, switch configs and network documentation. The data is useless if nothing can reach it.
Immutable retention
At least one copy that cannot be altered or deleted inside its retention window, by anyone.
Documented recovery plan
Written order of restoration, who does what, and how you operate while it happens.
An untested backup is an assumption
Backup jobs report success for years while writing data nobody can read back. Corrupted archives, missed volumes and databases captured mid-write all pass a completion check. The only meaningful test is a real restore to a real location, performed on a schedule and documented.
Ransomware changed the requirement
Modern attacks locate and encrypt backups before touching production, and they wait long enough that infected copies sit inside your retention window. Defending against that needs immutability, separate credentials, and enough retention depth to reach back past the intrusion.
When was your last verified restore?
If the answer is that backups are running, that is not the same thing. We can test one and show you what recovery actually looks like.
