History of Dev Ops at Zynga

When Zynga started out, it had to figure out what the right architecture for social gaming would be. Game teams were encouraged to innovate. As a result the role of dev ops was highly decentralized. After some time it realized that game architectures were converging, so it centralized the dev ops role. As part of that centralization it was able to make significant investments in tools and processes that enabled it to operate significantly faster and more robustly. It was also able to define the dev ops role. As a result: no Zynga game has experienced infrastructure related downtime during launch; it has reduced the time taken to prepare the game launch infrastructure; and it has improved the infrastructure efficiency. A direct consequence of that centralization was the evolution of its next generation architecture, the Zynga API, which allowed further centralization of dev ops. But the wheel continues to turn... and it has to once again look at how do it enables widespread infrastructure innovation while retaining the efficiencies and reliability of centralization.

In this talk, Kostadis Roussos, Chief Engineer at Zynga, will discuss that history and talk about the future.