Officials for the Canadian city of Calgary, Alberta, had planned to spend more than a year gradually migrating its remote sites to the city’s new multiprotocol label switching (MPLS) network. But, when the largest flood in the city’s modern history last summer threatened catastrophic network failure, they had only mere hours to save a network that included Calgary’s mission-critical communications.

As the storm bore down, a fast migration to the city’s new—but untested—MPLS network emerged as the only hope.

“We were getting ready to test the network, but we had not tested it yet—and then the flood hit,” David Basto, a project manager with the city of Calgary, said during an IWCE session on wireless backhaul.

A slow-moving, low-pressure system in southern Alberta put the new network to the ultimate test. The storm lingered for two weeks, mixing rain with an abundance of snow left over in the mountains from the winter. The resulting flood caused the river to swell and inundate downtown Calgary.  

“A test that is difficult to do is agility. How do you respond to an unplanned event or something that is completely unexpected?” Basto said. “Natural disasters have a tendency to do that. It will throw you off in way in a way you won’t even, can’t even imagine.”

Calgary was upgrading a legacy point-to-point network that included 230 remote sites, half of which were connected with city fiber coming from the city hall.

The problem? Calgary’s city hall, which housed the city’s network core, is located in a 100-year flood plain and such a once-in-a-century storm was happening.

An already precarious situation would have been worse, if the city had not moved its data center from city hall to a new emergency operations center in 2012. The timing of that move would prove fortuitous.

“They told us [that], in two and a half hours, we were losing city hall. Catastrophic failure on the network, right?” Basto said. “So, we had to come up with a plan very quickly to respond to this. All our 911 systems, all of our emergency systems are on this network.

“So we had to consider what to do at this point. Do we try to solve the power problem? Because basically it was a power issue … It was going to take a long time to restore everything.