December 11, 2012, 2:04 PM — During Hurricane Sandy this fall, The Weather Channel experienced its highest traffic ever. Normally the media company -- which spans television, desktop and mobile platforms -- supports about 90 million Web and mobile users a month. During Sandy, that jumped to 450 million -- nearly double the company's previous high for Web traffic.
Fortunately, The Weather Channel was prepared for the traffic surge. Landon Williams, VP of Platforms and Orchestration, and his IT team had recently architected the company's real-time radar mapping system to run on Amazon Web Services' Elastic Compute Cloud (EC2). On normal days, the mapping system runs on about 20 instances, but during Sandy it scaled up to run on 175 nodes in AWS's cloud.
A LOOK BACK: Top Networking and IT Stories of 2012
A LOOK AHEAD: 10 cloud predictions for 2013
CLOUD SHOWDOWN: Amazon vs. Rackspace (OpenStack) vs. Microsoft vs. Google
It's a classic use-case for the cloud: Variable and unpredictable peak demand for Web services is outsourced to a public cloud provider. It would have required a major investment for Williams to spin up a system across the company's on-premise and colocation environments to support the traffic load it experienced when Sandy ripped up the Eastern Seaboard, and then it would have likely gone mostly unused during other times. The alternative would have been not to have the compute horsepower to serve all the visitors to The Weather Channel's platforms, but Williams doesn't like that option either.
Adequate compute power for the application allows the radar maps to be updated as close to real-time as possible. Without the scaled-up, on-demand cloud resources users would have experienced lagging maps, delayed by seconds or maybe even minutes. "I want to be able to give people that visit our site the most up-to-date information possible," he says. In extreme circumstances -- think hurricanes, tornadoes, floods, lightning -- weather information can be a life-or-death matter.