You have backups — but how long will recovery take after a disaster, and how much data will be lost along the way? These are the questions disaster recovery (DR) answers: restoring IT systems after a serious failure, an attack or the loss of a site. And the answers are expressed in two figures: RTO and RPO.
Below: what they mean, how to set them and how to write a recovery plan that works.
RTO, RPO and MTD in plain words
NIST’s contingency planning guidance defines these terms as follows:
- RTO (Recovery Time Objective) — the overall length of time a system’s components can be in recovery before negatively impacting business processes. Put simply: how soon the system must be running again.
- RPO (Recovery Point Objective) — the point in time to which data must be recovered after an outage. Put simply: how much data you can afford to lose.
- MTD (Maximum Tolerable Downtime) — how long a business process can be disrupted without significant harm to the organisation.
A system’s RTO has to fit within the MTD of the process that depends on it. RPO determines how often you back up: with an RPO of one day, a daily backup is enough; with 15 minutes, you need replication or much more frequent copies.
Where to start: business impact analysis
NIST describes business impact analysis (BIA) as analysing operational functions and the effect a disruption might have on them. In practice:
- List business processes: sales, production, accounting, customer service.
- Identify the systems they rely on and the dependencies between those systems.
- Estimate the MTD of each process together with its owner.
- Set RTO and RPO for each system.
- Prioritise recovery.
An example of the result:
| System | RTO | RPO | How to achieve it |
|---|---|---|---|
| Accounting system | 4 hours | 1 hour | Replicating the virtual machine to a backup site |
| 8 hours | 4 hours | A cloud service or frequent backups | |
| File server | 1 day | 1 day | Daily backups under the 3-2-1 rule |
| Document archive | 1 week | 1 week | Weekly backups |
The lower the RTO and RPO, the more expensive they are to achieve. That is why they are set per system, not “everything, immediately”.
Recovery strategies
- Restoring from backups. The cheapest but the slowest: you need somewhere to restore to and time for the restore itself. The foundation is the 3-2-1 rule.
- A backup site. A “cold” site has premises and hardware still to be configured; a “warm” site has prepared infrastructure with not-quite-current data; a “hot” site has continuous replication and fast failover.
- Replication to the cloud. Virtual machines replicate to a public cloud and start there in a disaster — see “IaaS, PaaS and SaaS: what to choose for business” and “Moving servers to the EU: when it makes sense”.
- High availability. A cluster in which another node takes over the work of one that fails. This protects against hardware failure, but not against losing a site or ransomware — so it does not replace DR.
What a recovery plan needs
- Scenarios: a failed server or storage, loss of a site, ransomware, a long blackout.
- Roles and contacts: who decides to fail over to the backup site, who restores, who talks to customers. Contacts must also exist outside corporate email, which may be down too.
- Recovery order: first the network, directory services and DNS — see “Active Directory security: the basic settings” — then critical systems, then the rest.
- Where the backups are and how to reach them if the main network and accounts are unavailable.
- Criteria for returning to the main site.
NIST notes that the contingency plan is the first plan used to determine what happened and what to do, and for major disruptions it may point to the disaster recovery plan.
Ransomware is a separate scenario
After a ransomware attack, simply restoring backups is not enough:
- backups must be isolated or immutable, otherwise they get encrypted with everything else;
- before restoring, make sure the attacker is no longer in the network, or the attack will repeat;
- restore into a clean environment and change privileged account passwords.
First steps during an attack are on the “Security incident” page.
Testing
A plan that has never been tested is just a document. What to do regularly:
- test restores of individual systems from backups;
- tabletop exercises: the team walks through a scenario “on paper” and looks for gaps;
- failover to the backup site for the most critical systems;
- updating the plan after changes to the infrastructure.
NBU requirements
For insurers, credit unions, finance companies and pawnshops, NBU Resolution No. 143 makes cyber incident response planning part of contingency planning: the response plan must take business continuity documents into account or be part of the business continuity plan (paragraph 40). More in “Cyber incident response plan under NBU rules”.
Checklist
- Acceptable downtime is defined for key business processes
- RTO and RPO are set for each system
- The recovery strategy matches the RTO and RPO, not just the budget
- Backups are isolated or immutable
- There is a recovery plan with roles, contacts and recovery order
- Contacts and instructions are available without the corporate network
- Recovery is tested regularly
- The plan is updated after infrastructure changes
How Bereg-IT can help
We build backup: a preliminary backup audit, the 3-2-1 strategy, backup as a service in the cloud and on your own servers, recovery of files and applications, including AD, SharePoint, MS SQL and Oracle. A private cloud gives high availability: the role of a failed node moves to another. And for a backup site we help choose a cloud platform, including in EU data centres.
The first consultation is free. Send a request and an IT engineer will get in touch.
Sources
- NIST CSRC Glossary, definitions from NIST SP 800-34 Rev. 1: recovery time objective — https://csrc.nist.gov/glossary/term/recovery_time_objective; recovery point objective — https://csrc.nist.gov/glossary/term/recovery_point_objective; maximum tolerable downtime — https://csrc.nist.gov/glossary/term/maximum_tolerable_downtime; business impact analysis — https://csrc.nist.gov/glossary/term/business_impact_analysis; disaster recovery plan — https://csrc.nist.gov/glossary/term/disaster_recovery_plan
- NIST SP 800-34 Rev. 1, Contingency Planning Guide for Federal Information Systems (May 2010): https://csrc.nist.gov/pubs/sp/800/34/r1/upd1/final
- Resolution of the Board of the National Bank of Ukraine of 09.12.2025 No. 143, paragraph 40 of the Regulation (in Ukrainian): https://bank.gov.ua/admin_uploads/law/09122025_143.pdf