In 2010, I was part of the team which scaled the SMS microwork service of the Knowledge Generation Bureau for the television ad we ran during the American Super Bowl (along with others that day). That year, the Super Bowl had 106.5 million television viewers.
Since the kgb service is carried out by thousands of people working collaboratively on our web app, scaling our system was significantly more challenging than simply scaling a website or web application with limited collaboration.
My role involved:
Developing and coordinating distributed, cloud-based simulations of worker activity in order to:
Establish what our ordinary architecture was capable
Advise the business on possibilities for the ad campaign
Give the business information it could use to decide if it should approve the ad campaign
Provide the tech team with the information it needed to decide a scaling approach
Run simulated performance tests as part of the development and deployment process of scaling and testing our new architecture
Briefing our technology partners on the superbowl project, what our plans for scaling were, and discussing how to coordinate our efforts