Longer and bigger! Large-scale Internet outages are reaching increasingly impressive and worrisome levels. This is what millions of inhabitants of Grenoble, Besançon, Reims (Marne) and Strasbourg experienced again on April 27, after the sabotage of several sections of optical fiber.
Immediately, an investigation was opened for, in particular, “deterioration of property likely to harm the fundamental interests of the nation”. It’s because, in an increasingly connected world, being deprived of access to the Web quickly turns into a nightmare, for businesses and individuals alike.
ThousandEyes, an American company specializing in the identification of incidents on the network, has thus applied itself to drawing up a list of the biggest service outages that have occurred during the year 2021 around the world.
At the top comes the seven-hour Meta services (Facebook, Instagram, WhatsApp), which occurred on October 4 and which hit “hundreds of millions, if not billions of users”through the customers of the company’s partner companies.
The podium is completed by Fastly (a service that speeds up website loading times), with 85% of customers, including The worldthe New York Timesthe Guardian, GitHub, Twitch, Reddit and PayPal, on June 8, saw their websites made partially or totally unavailable for one hour; and by Amazon Web Services, the world champion of the cloud (dematerialized computing), whose successive blackouts on December 7 and 15 disrupted the proper functioning of platforms as popular as Netflix or Disney+, but also handicapped the services of the parent company (Amazon Prime, etc.).
In France, another breakdown marked the spirits, the one, linked to the fire, which ravaged the OVHcloud installations on the night of March 9 to 10, 2021.
Small number of actors
The enumeration is enough to highlight the variety of actors who can be at the origin of these failures, but also the multiplicity of causes, from the sabotage of equipment to accidental fires, by the way, and this is the most frequent cause. , by human intervention to update the software of this equipment, as in the case of Fastly. “The first reason for outages is the complexity of the Internet, with a tangle of interdependent solutions”notes Nicolas Chagny, the president of the association Internet Society France.
The same underlines that the extent of the breakdowns testifies less to a lack of reliability of current technologies than to their concentration in the hands of a small number of actors: “The cloud has accelerated the concentration on some very large operators, whose outages affect a large number of people”, he explains. In the field of cloud infrastructure, three players (Amazon at 33%, Microsoft Azure at 20% and Google at 10%) alone control more than 60% of the market and continue to grow at high speed by offering more and more services. additional services to their customers.
You have 46.44% of this article left to read. The following is for subscribers only.