What Is a Disaster Recovery (DR) Center? When to Choose DRaaS
Nanda Teknoloji Engineering Team · · 8 min read
In brief
A disaster recovery (DR) center is a second environment that takes over critical systems when the primary location becomes unavailable. DRaaS is renting this environment as a service; it suits organizations that do not want to build and operate their own second location or lack the capability to do so.
A disaster recovery (DR) center is a second environment that takes over critical systems when your primary location becomes unable to operate because of a fire, power outage, network failure, hardware fault, or cyberattack. It is not just “a backup sitting far away”: in addition to data, it covers server capacity, network connectivity, access routing, and the activation process. DRaaS (Disaster Recovery as a Service) is the model in which this environment is built and operated by a service provider. It is the typical choice for organizations that do not want to invest in their own second data center but still need to meet their RTO and RPO targets. This article covers what DR is, the standby models, and when DRaaS makes sense, using a concrete decision framework.
Are disaster recovery and backup the same thing?
No, and this difference leads to expensive surprises in practice. Backup preserves past points in time of the data. Disaster recovery, on the other hand, aims to take this data and turn it into a running service. No matter how solid your backup is, if the server, storage, network, and identity infrastructure to restore it is not ready, recovery can take days.
| Topic | Backup | Disaster recovery |
|---|---|---|
| Purpose | Being able to roll data back to a past point | Bringing the service back up in a second environment |
| What it protects | Data | Data, applications, network, access, and process |
| Typical scenario | Deleted file, corrupted database, ransomware | Loss of a location, major infrastructure failure |
| Time target | Usually hours | Minutes to days, depending on the target |
| Setup | Software and repository | Second environment, replication, runbook |
The two are not alternatives to each other but complements. In a ransomware scenario, the DR environment fed by replication may also receive the corrupted data, so an immutable backup that lets you go back in time is still needed.
Hot, warm, and cold DR models
It is common to divide DR environments into three classes by readiness level:
- Hot: The second environment runs continuously, and data is replicated synchronously or nearly synchronously. Takeover takes a very short time. It is the highest-cost model, because you pay for capacity continuously.
- Warm: The infrastructure is partly set up, critical data is replicated, and applications are started or scaled at the time of takeover. It offers medium cost and a medium recovery time.
- Cold: The second environment holds only backups and setup scripts; servers are prepared when needed. It is the lowest-cost model but has the longest recovery time.
Which model to choose depends not on technical taste but on the RTO/RPO set for each system. Rather than applying a single model to the whole environment, it is more economical to build a tiered approach: payment and order systems can be kept hot, ERP warm, and archive and test environments cold.
How does DRaaS work?
In DRaaS, the provider delivers the infrastructure at the second location, the replication software, and, often, the operation of the takeover process as a service. A typical flow is as follows:
- Production systems are replicated continuously or periodically to the provider’s environment.
- The replica is kept ready according to a defined recovery order (identity and network first, then the database, then the application).
- In the event of an outage or a declared disaster, systems are started in the provider’s environment.
- User traffic is redirected to the new environment.
- When the primary location is restored, failback is carried out in a planned way.
The critical point here is that what happens after takeover must also be designed: DNS and load routing, firewall rules, certificates, third-party integrations, and user access must be thought through from the start. If the provider follows a planning discipline of the kind described by NIST SP 800-34, these steps will be documented; for details, see NIST’s Contingency Planning Guide. There is also an entry for DRaaS with a short definition of the term.
When is DRaaS preferred?
| Situation | Own DR location | DRaaS |
|---|---|---|
| Capacity to invest in a second location | Yes | None, or not wanted |
| Team to operate DR | In place and experienced | Limited |
| Demand fluctuation | Fixed and predictable | Variable, flexible capacity needed |
| Need for control and customization | Very high | Medium |
| Speed of getting started | Slow (setup) | Fast |
| Regulatory or contractual location requirement | Met at your own location | The provider’s location must meet the requirement |
DRaaS typically stands out in the following situations:
- When the budget prefers a usage-based flexible model over a second environment that sits idle all the time.
- When the IT team is busy with daily operations and building a separate capability for DR operations is difficult.
- When the goal is to quickly bring certain systems under protection, and redesigning the whole environment is not necessary.
- When there are multiple locations or branches and a central, consistent recovery approach is wanted.
An own DR location may be more suitable in these cases: if very specialized hardware or high-volume data movement is required, if full control is mandatory, or if the organization already operates multiple data centers. In hybrid cloud structures, a mix of the two models is also common.
What should a DR plan include?
- Scope and ordering: Which systems will come up, and in what order?
- Decision and declaration process: When and by whom is a “disaster” declared?
- Roles and communication: Who does what, who is notified, who takes over the backups?
- Technical runbook: Step-by-step takeover and failback procedures.
- Dependencies: Identity, DNS, certificates, licenses, external service integrations.
- Test schedule: A plan is not considered valid unless it is tested. For the method, see our disaster recovery test checklist article.
Common mistakes
- Choosing the architecture without setting targets. First RTO/RPO, then the model.
- Leaving the DR environment different from production. A difference in versions or configuration produces errors at the moment of takeover.
- Protecting only the data. Without network, identity, and routing, data alone is not a service.
- Failback without a plan. Failback is often more complex than takeover.
- Not testing. A DR plan that has never been tried is an assumption.
- Leaving the provider’s responsibility unclear. In DRaaS, who does what must be clear in the contract and the runbook.
Example scenario: staged recovery with warm DR
A mid-sized manufacturing company runs its ERP system, file servers, and email infrastructure at its primary location. A long power and network outage occurs at the primary location. The organization has chosen the warm DR model: the ERP database is replicated at frequent intervals to the second environment, while the application servers wait as powered-off templates.
Takeover proceeds in this order: first the identity service and DNS, then the ERP database, the application servers, and finally user access. Because the file servers are of lower criticality, they are restored from backup by the end of the same day. What matters here is that the order and roles were written in advance; instead of making decisions in the crisis, the team follows the runbook. This scenario shows that disaster recovery is a matter of sequence and discipline more than technology.
Questions to ask the provider
If you are considering DRaaS, get clear answers to these questions at the proposal stage:
- Which workload types and platforms are supported?
- What is the replication method, and how is the achieved RPO monitored and reported?
- Is the takeover process automatic or manual; who triggers it and who operates it?
- What are your test rights; can you test without affecting production, and is test frequency limited?
- How are network, identity, and routing handled on the provider side?
- Who carries out the failback process, and does it incur extra cost?
- In which location is the data kept, and does it meet your residency needs under KVKK?
- Which measures do the service levels committed in the contract cover, and what happens in case of a breach?
Unless the answers to these questions are reflected in the contract and the runbook, they remain only in the sales presentation. Be sure to compare whether the service level commitments meet your own targets (RTO/RPO).
The difference between DR and high availability
The two are often confused. High availability (HA) aims to mask component failures (server, disk, network card) within a single location through automatic takeover. Disaster recovery covers major events such as the loss of an entire location. An HA cluster structure can reduce downtime to seconds in a server failure, but when the data center becomes unreachable, the cluster goes down with it. That is why the two approaches are planned together: HA for everyday failures, and DR for location-level disasters. Writing out in a table which layer covers which type of outage prevents false expectations about targets.
Decision summary
- DR is not a replacement for backup but a layer added on top of it that makes it possible to take over the service.
- The model (hot, warm, cold) is chosen according to the systems’ RTO/RPO targets.
- DRaaS is a fast and operable path for organizations with investment, team, and flexibility constraints.
- The plan must also cover dependencies and failback, and it must be tested regularly.
Moving forward with Nanda Teknoloji
Nanda Teknoloji carries out the discovery, design, migration, and operations steps in projects like this: classifying critical systems, choosing the appropriate DR model, documenting takeover and failback processes, and running drills. For a structure suited to your organization, you can review the disaster recovery center and backup solutions pages.