The public cloud wilts in the UK heatwave
Google and Oracle both confessed to data centre outages as temperatures hit 40C in the UK yesterday.
July 20, 2022
Google and Oracle both confessed to data centre outages as temperatures hit 40C in the UK yesterday.
The Google Cloud status update site refers to an ‘Incident affecting Google Compute Engine’, caused by a ‘cooling related failure in one of our buildings that hosts zone europe-west2-a for region europe-west2.’ Oracle also confirmed its ‘Cloud Infrastructure Compute UK South (London)’, among other services, was affected by a ‘cooling infrastructure’ issue in the associated data centre.
Both companies have, of course, been on the case to resolve the issues and seem to think everything was back to normal by the end of the day. But that still means users reliant on those public cloud services were in limbo for a few hours, unless they had redundancy with other providers or within their own private data centres.
While the temperatures were exceptional for the UK, there are plenty of other parts of the world where they’re a more common occurrence in the height of summer. It is surprising, therefore, that these multinational cloud giants have such inadequate capacity in the cooling systems that are so vital to the functioning of a data centre. We would imagine their customers, especially those directly affected by the outages, are having similar thoughts.
Every time one of the dominant US hyperscalers experience such an outage it begs the question of how wise it is for much of the business world to become increasingly reliant on them for almost everything. We don’t doubt the likes of Google and Oracle are doing their best to provide a robust and reliable service but they can’t control everything. What if the cooling failure had led to a more severe and prolonged outage?
Such a concentration of resource also raises concerns about the commercial power it grants those few who control it. Yesterday also saw the release of internal documents from Amazon, Google and Facebook which, according to a report from the Verge, reveal how they favoured their own products in order to stamp out competition. The report doesn’t identify any especially shocking abuses of power but it does remind us that they do happen. While these documents don’t directly concern the public cloud, they may speak to the culture of the companies concerned.
Another recent example is a class action suit taking place in the UK against Google and Apple for alleged abuse of their app store duopoly. The Google part of it was given the green light to proceed in the UK, according to a report by Foss Patents. It concerns the allegation that app store owners abuse their position by excluding all payment processing options other than their own and are thus able to impose excessive commissions on their captive market.
It is in the nature of all organisations, private and public, to be self-serving; the incentive structures almost guarantee it. Which is why outages experienced by one of the public cloud oligopoly are so concerning. Could they have spent more on the cooling systems in question and, if so, were they persuaded from doing so by short term commercial considerations? Right now there seems no way of finding out, nor of regulating these companies we increasingly rely on for so much. Surely that has to change.
Get the latest news straight to your inbox. Register for the Telecoms.com newsletter here.
About the Author
You May Also Like