Mycelium 365 — Managed IT, Microsoft 365 and Azure across Australia, New Zealand and the United States
    Resilience

    Disaster Recovery (DR)

    Disaster recovery is the capability to restore IT systems and data after a disruptive event such as ransomware, hardware failure, cloud region outage or physical damage. It comprises documented plans, replicated infrastructure or backups, defined recovery objectives, and regular tested exercises.

    What Disaster Recovery (DR) means in practice

    Disaster recovery is the technology component of business continuity. Continuity planning asks how the organisation keeps operating; disaster recovery asks how the systems come back. The two are frequently conflated, with the result that a technically successful restore lands in an organisation that has no process for the period before it completes.

    A credible DR capability has four parts. A documented plan naming who declares a disaster, who executes, and in what order systems are recovered. A dependency-ordered recovery sequence, because restoring an application before identity, DNS and network connectivity achieves nothing. Protected copies of data that an attacker with domain administrator rights cannot delete — immutable, offline or in a separate security boundary. And a testing regime that proves the whole thing under time pressure.

    Ransomware changed the design assumptions. Historical DR planned for single-site loss, with the surviving site intact. A ransomware event compromises the environment itself, including backup servers and often the backup catalogue, and may have been present for weeks before detonation. That drives requirements for immutability, credential separation between production and backup, and retention long enough to recover from before the intrusion began rather than after.

    Cloud does not remove the obligation. Azure provides regional resilience, availability zones, geo-redundant storage and Azure Site Recovery, but the customer still chooses and configures them, and shared responsibility places recovery of customer data with the customer. A single-region deployment with locally redundant storage is a design decision, not a default someone else is managing.

    The habit that separates organisations that recover from those that do not is exercising. An annual test that restores a real workload, times it, and produces a list of what went wrong is worth more than any amount of documentation, and is the evidence insurers and regulated clients now ask to see.

    How we help with this

    Related terms

    Back to the full glossary

    Ready to simplify and secure your technology?

    Book a free, no-obligation Discovery Call to talk through your Microsoft 365, Azure, security, or support needs — no sales pitch, just a straight conversation.

    We respond to every enquiry within 4 business hours. Monday to Friday, 7am–7pm AEST.