Pay Now
Book a security review

NIST CSF 2.0 · RECOVER

Disaster Recovery

Almost every business has a number in mind for how quickly it could recover, and almost none of them have tested it. Backup is about the copy existing. Disaster recovery is about how long it takes to be trading again, which depends far more on the sequence, the dependencies and the people than on the data itself.

NIST CSF 2.0 · RECOVERRTO and RPO agreedSized to actually meet themProven by a real failover

The problem

The recovery time in your head has never been measured

Ask what would happen if the main system were gone this morning and you will get a confident answer. We would restore from backup. It would take most of a day.

What that answer usually leaves out is everything except the data copy. Somewhere to restore onto. The order systems have to come up in, because the application will not start without the database and the database will not start without the network. The licence keys. The credentials for the one system nobody uses often. Whether the person who knows how to do this is contactable.

That is why measured recovery times are so consistently longer than assumed ones. The data is rarely the constraint. The sequence is.

Two numbers, set per system

How much time, and how much work

Applying one target to everything is what makes disaster recovery unaffordable. Most businesses have two or three systems that need to be fast and a long tail that does not.

RTO

Recovery time objective

How long the system can be unavailable before the consequences become serious. Set by what the business actually tolerates - an hour for the system that takes orders, a day for the one that produces reports. Every reduction in this number costs money, so it is worth being honest.

RPO

Recovery point objective

How much work you can afford to lose, measured in time. Fifteen minutes means replication; twenty-four hours means a nightly backup is sufficient. This one is frequently set by how much re-keying your team could realistically do.

How we build it

From the business impact backwards

STEP 1

Work out what each hour costs

Per system, in revenue, in staff time and in obligations you would breach. This is the conversation that sets the targets, and it is a business conversation rather than a technical one.

STEP 2

Map the dependencies

What has to be running before each system can start. Nearly every recovery plan we review is missing at least one dependency, and it is discovered mid-recovery when it is most expensive.

STEP 3

Design to the target

Replication for the systems that need to be back within an hour, standard restore for the rest. Paying replication prices for the whole business is how the budget disappears on the systems that mattered least.

STEP 4

Write the runbook

Step by step, in an order somebody can follow under pressure at three in the morning. Held somewhere that does not depend on the systems being recovered, which is a mistake made surprisingly often.

STEP 5

Fail over for real

A genuine failover test, timed against the target. Until that has been done, the recovery time is a design intention rather than a measured fact.

Be honest about the options

What each level of investment actually buys

  Backup and rebuild Warm standby Continuous replication
Typical recovery time One to five days Two to eight hours Minutes to an hour
Typical data loss Up to a day Up to a few hours Minutes
Ongoing cost Lowest Moderate Highest
Somewhere to restore onto Has to be found Ready Running
Tested how Restore test Failover test Failover test
Sensible when Days of downtime are survivable Most mid-market businesses Downtime stops revenue immediately

Questions

What businesses ask about Disaster Recovery

Short answers to what gets asked most. If yours is not here, put it to us directly and you will get the same answer in writing.

What is the difference between this and backup?

Backup answers whether the data still exists. Disaster recovery answers how long until people can work again. A business can have flawless backups and still be down for a week because there is nowhere to restore onto, no documented sequence, and one person who knows how it all fits together.

How do we choose a recovery time objective?

By costing the downtime rather than by picking a number that sounds responsible. Work out what an hour of that system being unavailable actually costs, then compare it against what shortening the recovery time costs per year. Most businesses discover they have one or two systems that justify real investment and a long tail that comfortably does not.

Does this work if we are entirely in the cloud?

The failure modes change rather than disappearing. A Microsoft outage is genuinely rare and largely outside your control. What is far more common is a tenant-level problem - a compromise, a mistaken bulk deletion, a misapplied policy - and recovering from that needs a plan and a copy of your data outside the tenant. Being cloud-based removes the hardware failure, not the incident.

How often should we test?

A full failover annually, and a partial test twice a year. Also after any significant change, because the most common reason a recovery fails is that the environment moved on and the runbook did not. An untested plan degrades quietly, and the degradation is invisible until the day it matters.

How it is bought

Included in a managed plan

This runs as part of a monthly managed plan - set up, operated and reported on by us, priced per person per month with Microsoft licences quoted separately.

Book a security review

Find out exactly where you stand, before someone else does

We will measure what you have against NIST CSF 2.0, tell you the three things worth fixing first, and put it in writing. No charge, and no obligation to buy anything afterwards.