Disaster recovery and business continuity for businesses

Backups answer the question “did we lose the data?”. Disaster recovery answers “when are we back to work?”. These are different questions, with different answers. A company can have perfect copies and be out of action for three days because nobody knows where to restore them, in what order, or with which server. DataRoad designs, maintains and tests disaster recovery plans for companies that cannot afford downtime, and writes them with two numbers on the first page: how much can be lost (RPO) and how quickly everything needs to be back up and running (RTO). This includes recovering data from backups whenever necessary — not just in a total disaster scenario.

Backup and disaster recovery are not the same thing

As managed backups they are the foundation; the disaster recovery plan is what is built on top of them.

The two numbers that matter: RPO and RTO

RPO — how much can be lost

O Recovery Point Objective it is the maximum amount of work that the business is willing to lose, measured in time. An RPO of 24 hours means that, in the worst case, one day's worth of work is lost; an RPO of 15 minutes requires almost continuous replication. There is no single correct value — there is only the value the business is willing to pay. A clinic that books 80 appointments a day and an accountancy firm at month-end have very different RPOs.

RTO — how long until you go back to work

O Recovery Time Objective It is the time between the disaster and the moment when critical systems are up and running again. It determines the architecture: with a two-day RTO, it is enough to have somewhere to restore to; with a two-hour RTO, you need replicas ready to boot up. The RTO is written into the contract and is measured during every test.

What does a DataRoad disaster recovery plan include

• Impact assessment. Which systems stop the business if they go down, for how long is it tolerable, and how much does each hour of downtime cost. This is where the RPO and RTO of each system come from — they do not have to be the same for all of them.
• Backups using the 3-2-1 rule. Three copies, on two different media, one off-site and, for critical systems, an immutable version that not even a compromised administrator can delete.
• Replicas of critical servers. In a data centre in Portugal or in the cloud, according to the required RTO, ready to start without waiting for new hardware.
• Communications plan. How does the company continue to answer phones and receive email while the main infrastructure is down — includes the switchboard and Microsoft 365.
• Written, step-by-step procedure. Who decides to declare the disaster, who is notified, in what order systems are recovered, what passwords and contacts are needed. Kept outside the systems you may need to recover.
• Tests scheduled. At least one full test per year and quarterly partial tests. Each test produces a report showing the actual recovery time compared with the agreed RTO and the corrections to be made.
• A return to normality. The plan does not end when the company is working on the replica; it includes the process of returning to the main infrastructure without losing the work carried out in the meantime.

The most frequent case: ransomware

In Portugal, the most common disaster is no longer fire or flood — it is an encrypted file that spreads through the network overnight. A disaster recovery plan prepared for ransomware has three characteristics that a generic plan does not: immutable copies, which the attacker cannot encrypt or delete even with administrator credentials; a network isolation procedure, to prevent the replica from becoming infected upon boot; and a securely identified restoration point prior to the infection, because restoring yesterday's copy may mean restoring the ransomware itself. In cases where we are called without a prior plan, the job is one of data recovery — slower, more expensive and with no guarantees.

Who it makes sense for

For any business where a single day without systems costs more than the annual subscription fee. The clients who need it most are hotels whose bookings and check-ins depend on the management system, clinics with appointment schedules and medical imaging, law firms and accountants with legal deadlines, embassies and organisations with availability requirements, and logistics firms where invoicing and dispatches grind to a halt simultaneously. Smaller businesses benefit from a simplified version: off-site backups, a one-page written procedure and one test per year.

Inside IT Unlimited

Managed backups and monitoring are included in all plans IT Unlimited. The comprehensive disaster recovery plan — including replicas, a written procedure and recovery tests — is included in the IT Unlimited Premium 365, the plan for businesses that cannot afford downtime, and can be contracted as a project for clients on the other plans.

How we got started

We start with the impact analysis, together with company management and not just with IT: it is management that knows how much an hour of downtime costs. From there, the RPO and RTO are established for each system, the solution is designed and the procedure is written. The first complete test takes place within the first three months. If you already have backups and just want to know if you would be able to recover from them, that test can be carried out in isolation, as consultancy work, with a written report.

A real-life case: server, UPS and backups in a local authority

The Union of Parishes of Massamá and Monte Abraão awarded to DataRoad, for public contract in 2022, the installation and configuration of a new server, a UPS system and corporate backups — the three foundational layers of any business continuity plan: processing capacity, power that can withstand a failure, and recoverable data. In a local authority that provides services to the public, a system outage is not just a cost; it means services cannot be provided; the ongoing management contract, Renewed through a tender, it combines monitoring and 24×7 support (Premium 365) with this base so that incidents can be detected before they affect the front desk.

What is put in writing

A disaster recovery plan that isn't written down does not exist. The document that the client receives includes:

• The impact analysis, system by system, with the RPO and RTO agreed for each one and the estimated cost of one hour of downtime.
• The architecture of copies and replicas: where each copy is located, how often it is created, which one is immutable, and how far back it is possible to go.
• The activation procedurewho declares the disaster, who is notified, the system recovery order and the necessary contacts and credentials, stored outside the infrastructure.
• The communications plan — how the company continues to answer phones and receive emails while the core systems are down.
• The test reports, including the actual recovery time relative to the RTO and the corrections decided upon after each test.

Frequently Asked Questions

Disaster recovery is one of IT services for businesses by DataRoad, in collaboration with the managed backups, a 24×7 monitoring (Premium 365) and a cybersecurity.

Where we operate

We design and test disaster recovery plans for clients across the country; our on-site support mainly covers Greater Lisbon. We have dedicated pages for Lisbon, Cascais, Oeiras, Sintra, Leiria, Coimbra e Porto. See all areas of expertise.

Contact us

Se quer saber em quanto tempo a sua empresa voltaria a trabalhar depois de um desastre — e não em quanto tempo gostaria — a primeira conversa é sem compromisso. Ligue +351 211 459 950 (weekdays, 09:30–13:00 and 14:00–18:30) or write to sales@dataroad.pt.

See also: NIS2: what changes for companies in Portugal

Practical guides on the DataRoad blog: Outage: could your company handle 8 hours of downtime? · Has your backup already been tested? The 30-minute test · How much does an hour of downtime cost