Disaster Recovery: what is it and why does a business need it

Disaster Recovery (DR) – these are not backups or "what to restore later". This is a pre-prepared plan that helps quickly launch the system after a disaster with the least loss of data and time. If you don't have a plan and replication, you don't have fault tolerance.

Disaster Recovery is a collection of:

• technical solutions,
• regulations,
• people and SLA,

which ensure the restoration of IT systems in case of failures: data center failure, fire, ransomware, human factor, sanctions risks.

How DRaaS works

• Your infrastructure replicated to a backup site (most often the cloud).
• The data is synchronized on a schedule or almost in real time.
• In case of an accident, it is executed failover — running a copy of the system.
• Users continue to work with minimal downtime.

Ways to organize Disaster Recovery

In 80% of cases, DRaaS — the best option in terms of price and recovery speed.

How Disaster Recovery differs from backups

A key point that is often ignored.

Fact: backups are part of DR, but never a substitute.

Key Disaster Recovery Parameters

• RTO — acceptable downtime.
• RPO — acceptable data loss.

Example:

What is the Disaster Recovery Plan?

No plan, no recovery.

What does the DRP plan include?

• Composition of the DR team and areas of responsibility
• Assessment of external and internal risks
• Mission-critical business processes
• RTO and RPO for each system
• Accident scenarios and procedures
• SLA with the provider
• Regular failover tests

Important: the Disaster Recovery plan without testing is paper.

Phased disaster recovery

• Incident detection
• Making a failover decision
• Launching backup infrastructure
• Information integrity check
• Switching users
• Analysis and return to the main environment (failback)

Disaster Recovery Parallel Infrastructure: where to keep a reserve

• In the cloud — faster startup, less CAPEX
• In the second data center, it is more expensive, but it is suitable for regulators.
• A hybrid is often the best option for a large business.

Who needs Disaster Recovery

Necessarily, if:

• downtime = direct financial losses;
• there are online services;
• regulatory or customer requirements;
• The business is open 24/7.

Examples of industries:

• Banking and finance,
• e-commerce,
• SaaS and IT companies,
• service companies.

A real case from the practice of a hosting provider

Task: to ensure the continuous operation of an e-commerce project with a turnover of ~30 million ? per month and peak loads during the sales season.

The initial situation

The main infrastructure was located in one data center:

• 4 VMs (web, app, database, queue),
• PostgreSQL + Redis,
• daily backups.

Technically, there is a backup, but there was no plan.

The actual RTO is several hours, and this would be critical for business.

The incident

As a result of a failure on the storage side, mainly the data center:

• the database has become unavailable,
• the site and API stopped responding,
• restoring from backups would take 6-8 hours.

For the client, it meant:

• direct sales losses,
• the load on the call center,
• reputational risks.

Disaster Recovery Plan Implementation

We have implemented DRaaS with the following architecture:

Replication of virtual machines to a cloud backup platform.

Asynchronous DATABASE replication with RPO takes 5 minutes.

Prepared Disaster Recovery Plan:

• accident scenario,
• responsible persons,
• the failover order.

Configured automatic trigger infrastructure startup.

Parameters after Disaster Recovery implementation

Repeat incident (after 4 months)

There was a network incident on the side of the main provider:

• The Disaster Recovery Plan was launched according to the regulations,
• the backup infrastructure went up automatically,
• users noticed only a short-term degradation.

Fact: Business did not stop, sales continued, and the SLA was met.

This case clearly shows the key point:

• Backups save data. Disaster Recovery saves the business.
• It pays off in the first incident, especially where downtime is measured in money rather than abstract "inconveniences."

FAQ:

Result

Disaster Recovery (DR) — this is a tool for the continuous operation of the company, which either exists and is tested, or it does not exist at all.

If the plan has not been tested, consider that it does not exist.