To stay agile, Phoenix Labs runs all its game servers in containers on the Google Cloud Platform (GCP). If the game hiccuped, players would know because they would fail to get into a match or face a slowdown in gameplay - something that could easily get them killed and very angry. They had to guess at the server capacity required, and the underlying cloud-based service had to work in an uninterrupted way, handling the real-time combat. But the teams had to plan for a number of scenarios without knowing exactly how many players would show up at any given time. “We had plenty of things that we wanted to develop and add” to the game, which is a kind of living experience that can be changed and improved day by day, Beaumont said.īoth Beaumont and Rayan said it is impossible to predict if a game will go viral. The game developers offload the multiplayer server operations to Google and its vast data centers, and the developers concentrate on making the game fun and as bug-free as possible. This kind of game is increasingly common, but each job is demanding, Rayan said.
With each beast killed, the players earned new weapons and armor. It was a triple-A quality game, with good graphics that immersed players in the fantasy of hunting down beasts known as Behemoths in small groups. The role-playing game wasn’t a lightweight title. And that four-player experience is then repeated over and over across millions of players, all of whom have a unique gaming experience.Īs the demand spikes hit and Phoenix Labs was totally surprised at the influx of players, the cloud did its job. Kubernetes allows Phoenix Labs to encapsulate what is supposed to happen in a game when four players download the game and play it together over online connections. In an interview with GamesBeat, Phoenix Labs vice president of technology Simon Beaumont and Sunil Rayan, managing director of Google Cloud for games, said that the key to scaling it up to millions of players in a matter of days was containerization, or standardizing the game so it could run on any servers using Google’s Kubernetes engine.
0 kommentar(er)
