Google launched Meet in 2017 as an enterprise-centric alternative to its Hangouts chat service. (Google is constantly phasing out Hangouts and making it part of its usual messaging platform strategy, to connect users to meats and chats.) As coronovirus spread and more countries ordered orders to stay at home Released, people worked for video chat services. And to check up on family and friends. Google saw meats grow by 30-fold in the early months of the epidemic; Soon, the service was hosting 100 million meeting participants each day. this is too much.
Amidst all the profound changes people have made in response to the COVID-19, the infrastructure that went through the Internet experienced a change in usage patterns, as well, as people traded office time for domestic segregation. The companies handling these systems have been able to manage the new needs of most users. “You essentially took the peak and prolonged it by the day,” says Ben Trenier Sloss, Google’s vice president of engineering. “Use went up, but it was mostly that usage looked like the peak of the day, rather than that peaks went up dramatically.” However, some services saw usage spikes beyond normal.
Google prepares for emergencies on a regular basis through its disaster and incident response tests or DIRT. In these exercises, around 10,000 employees at a time will simulate localized natural disaster to Godzilla attack to deal with any kind of crisis. The Kovid-19 epidemic, however, exceeded the company’s most dramatic scenarios.
“We usually simulated a regional-level event,” Traynor Sloss says. “I’d never done DIRT for a global event, if I’ve been honest, because it was unlikely.” There was also a practical concern: making fun of an incident with worldwide impact would risk undermining the experiences of real Google users, a cardinal sin in the world of DIRT.
All of which meant Shehevitz, who led the incident response to Google Meet, and the teams involved had to find things on the fly. In particular it became clear that they were taking in far more new users than their most ambitious initial estimates.
“Initially, we started a plan to double our footprint, which is already huge. This is not a normal growth curve. We soon realize that was not going to happen enough, ”says Shehewitz. “We kept trying to make progress on building more runways … so that we would have time to work out solutions if things would arise on a long time horizon instead of waking up every day and like, what is on fire today?”
The complaint of this challenge was that the Google engineers involved in the response were working from home themselves, spread over four offices in three countries. “All those people who have worked on it — and it’s a large number of teams — even people working in the same place have never really been in a room together since it started , ”Says Scheewitz, based in Zurich, Switzerland. On a technical level that proved to be quite manageable; As you might think, Google prioritizes web-based tools that can be accessed from anywhere. But to establish redundancy for more than just bandwidth, 24-hour-day operation needs to be coordinated remotely. In a blog post detailing the response, Shehevitz described how everyone in the response role of an incident was assigned a “standby”, basically a sensible person who could step in when the principal was ill or took time off. could. (Particularly prudent measures during the global health crisis.)