Technology news and discussion on TalkFuse.
Control fault prevented backup generators starting at Google’s London cloud site
Google says a power outage at its London cloud facility on 20 October was worsened by a fault that stopped emergency generators starting automatically. The disruption affected services used by organisations including Lloyds, HSBC, Royal Mail and Revolut, and exposed a weakness in the site’s backup power controls.
An underground cable serving power feed A failed while feed B was disconnected for repairs after problems during a generator test. Although uninterruptible power supplies kept the facility running for about five minutes, a voltage-sensing flaw and misconfigured control system did not send the generators a start signal. Google says it has reprogrammed the controls, tested changes to the electrical setup and plans to improve incident notifications and infrastructure testing.
In short
- A London power cable failed while the second feed was under repair.
- Google’s backup generators did not start because of a control system flaw.
- Google says it has changed and tested the power setup.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
The incident shows why cloud providers should be held to high standards: organisations depend on their infrastructure for banking, communications and other essential services, so backup systems and timely warnings are central to resilience. The outage exposed a preventable weakness in generator controls while one power feed was unavailable, making stronger testing, safeguards and disclosure reasonable expectations.
The case against
A single equipment and control failure does not by itself show that the provider’s overall approach to resilience is inadequate; complex infrastructure can fail despite testing and backup systems. Google has described specific corrective steps, and a fair assessment should account for how quickly they reduce the risk and improve notifications rather than infer broader negligence from this incident alone.
Explained simply
Cloud data centres house computers and storage that organisations use to run websites, apps and other online services. They need a steady electricity supply: backup generators are meant to take over if mains power fails, while uninterruptible power supplies provide a short bridge between the two.
The London facility involved is operated by Google, which provides cloud services to businesses and public bodies. Organisations including Lloyds, HSBC, Royal Mail and Revolut use services affected by the outage, so a problem at one site can disrupt many familiar services at once.
The incident highlights how backup systems depend on both equipment and controls that detect a power failure and tell generators to start. Understanding those links matters because a backup system may not protect a facility if its components are unavailable or fail to respond as intended.
Read the full article at the source →
Originally published by The Register as “When datacenters lose power, generators should kick in. That didn’t happen for Google in London”.