- Market
- Saudi Arabia
- Written for
- CIO, COO
- Decision
- Prove the business can recover within targets it can actually rely on
- Languages
- EN · AR
The targets
The two numbers the business sets before engineering designs anything
A recovery design starts from what the business can tolerate, not from a technology. These are the conditions a plan is judged against, and each is a commitment before it is an architecture.
A recovery point per system
How much data the business can afford to lose sets how often the system is captured — minutes for a ledger, a day for a reporting store.
What it protects: A loss window the business chose, not one it discovers.
A recovery time per system
How long the business can be without the system sets how the restore is engineered — a warm standby, a rebuild from backup, or something between.
What it protects: An outage bounded by design rather than by luck.
An off-site, isolated copy
The recovery copy lives beyond the blast radius of the primary, so the event that takes production down cannot take the backup with it.
What it protects: A copy that survives the failure it exists for.
A restore proven by drill
The plan is exercised under realistic failure conditions and timed, so the targets are demonstrated rather than promised.
What it protects: A capability the business has seen work, not a document.
Earning the word plan
How a recovery plan earns the word plan through rehearsal
A plan on paper is a proposal until it has been run. The steps that turn it into a capability are the ones most programmes skip, because they are the ones that expose whether it works.
Set the targets with the business
Each system gets a recovery point and time the business signed off, because the cost of protection is decided by what its loss costs — and not everything deserves the same. Agreed targets, tiered by what each system is worth.
Build the restore path
The backups, the off-site copy and the rebuild steps are engineered to meet the target, including the keys and dependencies a restore silently needs. A documented path from failure back to service.
Drill it under failure conditions
The restore is run as if the primary were gone, timed and observed, so the gaps — a missing dependency, a key nobody kept, a step that takes far longer than assumed — surface in a rehearsal, not an incident. A measured recovery time and a list of what broke.
Close the gap and re-drill
The drill result is measured against the target; where it misses, the path is fixed and rehearsed again, so the plan converges on a recovery the business has actually watched succeed. A plan validated against its own target.
The silent failure
The backup that completed every night and restored nothing
The most common disaster recovery failure is not the absence of backups; it is backups that ran successfully for years and could not be restored when it mattered. The job reported success every night, so nobody looked. Then a real failure arrived and the restore missed a database that was never in scope, or needed an encryption key that left with an employee, or took eleven hours against a target of two because the process had never been timed. The backup was real; the recovery was imaginary, and the gap between them was only ever visible in a drill nobody ran.
That is why we treat the drill as the deliverable, not the backup. A backup is a means; the thing the business is buying is the ability to recover within a target it can rely on, and that ability does not exist until it has been demonstrated. The uncomfortable version of this conversation happens in a scheduled rehearsal, where a missed target costs a re-run. The comfortable version never happens — the gap is discovered during the outage, when it costs the business itself.
Our role
What we design and prove and what your operations team runs on the day
- 01
Targets before technology
We help the business set a recovery point and time per system, tiered by what each loss costs, because a single blanket target either over-protects the trivial or under-protects the critical.
- 02
A restore that has been run
We engineer the recovery path and then drill it under failure conditions, so the target is a measured result rather than an estimate, and the dependencies a restore quietly needs are found in rehearsal.
- 03
A runbook your team owns
We leave a documented, rehearsed procedure your operations team can execute on the day without us, and a drill cadence that keeps the plan honest as the estate changes.
What each system is worth, and the targets the business will commit to, are yours to set; we build the restore path, prove it in a drill, and leave a runbook your team can run — so recovery is a rehearsed capability, not a document filed and forgotten.
Common questions
What buyers ask
What is the difference between recovery time and recovery point, and who sets them?+
Recovery point is how much data you can afford to lose — the gap between the last capture and the failure — and it sets how often a system is backed up. Recovery time is how long you can afford to be without the system, and it sets how the restore is engineered, from a warm standby that switches over in minutes to a rebuild from backup that takes hours. Both are business decisions, not technical ones: they are a statement of what an outage and a data loss cost the organisation. Engineering’s job is to design a recovery that meets the numbers the business sets and to prove it does. A supplier who quotes you targets without asking what your downtime costs has skipped the only part that matters.
We take nightly backups. Isn’t that already disaster recovery?+
Backups are a necessary part of disaster recovery, but on their own they are not it. A backup is only useful if it can be restored within your recovery time, includes everything the system actually needs, and survives the event that took production down. Many organisations discover during a real failure that their nightly backup missed a dependency, needed a key nobody had, or takes far longer to restore than the business can tolerate. The difference between a backup and disaster recovery is the drill: a recovery you have run under failure conditions and timed against a target. Until that has happened, you have a backup and a hope, not a plan.
How often should a recovery plan actually be tested?+
Often enough that it is never far behind the system it protects, which in practice means on a regular cadence and again after any significant change to the estate. A plan drills successfully once and then drifts: a new database is added and never included, a dependency changes, a key rotates. Each of those quietly invalidates a plan that was proven last year. We set a cadence proportional to how fast the estate changes and how critical the system is, and we re-drill after major changes rather than trusting that a plan proven once stays true. A recovery capability is not a milestone you pass; it is a property you maintain.
In this cluster
Next step
Scope it with us
A recovery plan that states the target per system, describes the restore path, and is validated by a drill you watched succeed — so continuity is a rehearsed capability rather than an assumption.

