On 20 October, one of the three zones in Google Cloud's europe-west2 region in London lost power shortly after 7am UK time. According to engineers, backup generators did not engage at first. Those affected included Revolut, Lloyds, HSBC, Barclays, Royal Mail, DPD and the video platform used by UK courts. Most services recovered by the afternoon, and Google said it would publish an analysis of the incident.

What it means: The incident is a practical lesson in architecture. Google has only one cloud region in the UK, so customers that had built their systems in a single zone, or without a plan to move to another region, had nowhere to go. The fact that banks, delivery firms and public services were hit at the same time shows how concentrated critical infrastructure has become among a few providers. For the European financial sector, the DORA regulation demands exactly what was missing here: mapping of third-party ICT providers, resilience testing and exit plans. Supervisors will want to see whether banks did what they had said they would. For smaller businesses, including in Greece, the message is simpler: a payment or ordering system needs at least a second availability zone and a tested recovery procedure. The generator failure also shows that the physical infrastructure of data centres matters as much as software. Until Google publishes its official analysis, it is unclear why the other two zones did not fully absorb the load, and that is the most worrying question for anyone relying on multiple zones.

Read the original on Computer Weekly

Related service: Custom software

Related stories