For Microsoft, the best part of the epic snowstorms that hammered its headquarters for nine straight days is that the collaboration technology that employees used to work from home was as seamless as if they were still in the office.
While the 50,000 employees who live in Redmond, Washington were blocked in by roughly two feet of snow that fell from February 4th-12th, the Microsoft Teams platform they used to work from home didn’t flinch despite taking on a much heavier workload than usual. The storms that battered the Northwest were the first large-scale test for remote workers at Microsoft since the company began moving all its communications from Skype for Business to Microsoft Teams in the fall.
“The big win here is Teams stayed consistent—we didn’t see a big difference from before the storms even though 90 percent of our employees were working from home,” says Dan Babb, a senior service engineering manager in Microsoft Digital. “Our metrics stayed pretty much the same across all the different kinds of internet connections that our employees use at home—that’s pretty amazing compared to where we used to be.”
The technology mostly disappeared into the background – it just worked.
“The environment scaled up,” Babb says. “All of the employees who were working from home used their own home network connections, which are never going to be as good as the connection we have in our corporate buildings. The metrics we focus on did have a slight dip—which was to be expected—but still remained green.”
To geek out for a minute, Teams performed even better than you would think, says Jonathan Clare, also a senior service engineer in Microsoft Digital.
“Most people working at home are on WiFi,” he says, explaining that the bulk of the 50,000 employees who live in the Puget Sound region were working from home, seriously taxing the private internet services they were using to connect. “Despite that, Teams performed very similar to how it would if everyone was working in their offices—that we had that kind of parity is remarkable.”
Babb attributes the positive experience to the massively scalable Teams’ micro-service architecture that runs natively in the cloud.
Teams isn’t without challenges, but those are mainly around refining the user experience.
“When I talk, I can be clearly understood,” Babb says. “The chance of having any type of audio distortion has greatly reduced—for example, I don’t sound like I have robot voice, and my voice doesn’t sound like it’s being cutoff.” Teams optimizes the audio and video experience in limited bandwidth conditions, greatly improving the call quality experience.
The biggest test for Teams during the snowstorms was meetings, because the impact of having everyone call in at the same time during a weather event is exponential.
“When everyone is in the office, you can have 10 people attend a meeting in a conference room—you can aggregate everyone in the room down to one person joining a meeting,” Babb says. “When everyone calls in from home, the level of complexity and the scale of the load is exponentially higher because all 10 of those people have to join on their own.”
During the storm, Teams worked fine, despite everyone at the company joining all their meetings from home, all at the same time. “The service didn’t bat an eye,” Babb says.
Some employees anecdotally reported having a few calls drop during the weather event, but there’s no way around that when everyone in the region is dealing with the same storm system that caused intermittent power outages and congested public internet services, Clare says.
Clare and Babb say they’re happy with how well Teams held up in its first big test at Microsoft.
