More than two decades ago, People Can Fly made a name for itself as the team behind a throwback, frenetic first-person shooter set in the realm between heaven and hell.
In the twenty years since Painkiller's launch, they are growing into a global publisher and have continued to expand, with more than seven hundred and sixty employees worldwide.
But recently, People Can Fly found itself in need of a backend engine that could both support the eclectic mix of games they were working on, but also serve as a robust tool for future projects.
What followed was an intensive vetting process that put a broad swath of backend engines under the microscope, meticulously examining them and eventually even testing several on current projects.
The company was looking for a backend engine that allowed its products to remain agile in their game development, maintaining the power to make decisions and changes at the studio level.
The result—the decision to use the Pragma Engine on all of People Can Fly's games-as-a-service projects—is an example of pragmatic research and future-proofing an increasingly important element of game design.
Lots of games, one problem
People Can Fly developed many games since Painkiller put them on the map in 2004. That includes working on two Gears of War titles, co-developing Bulletstorm, helping with Fortnite, and—in 2021—releasing Outriders through Square Enix.
By the time Clifford Roche joined the developer in 2022, the company hadn't yet locked in a backend for online services that they were happy with.
Roche and several other people at the company started casting about for existing solutions and providers.
"We knew that there was value in making sure that whatever we were building for these products was going to be shareable as much as possible," Roche said. "We weren't going to try to reinvent the same thing for each title."
That idea launched the company into a deep vetting period that examined current backend solutions on the market.
Vetting
Starting with a blank slate, People Can Fly created a massive spreadsheet that compared each provider by their strengths, weaknesses, and the backend features they would need across their multiple games.
The primary requirement was that whatever backend used would allow People Can Fly to go in and change things on the fly.
"That was sort of a safety net for whatever other technology we put in place," Roche said.
Key personnel from People Can Fly met with companies big and small to discuss their tech and how it could be applied to their in-development games.
"We cast a wide net to figure out what technology was there, and then we started doing research," Roche said. "Some key needs surfaced early on, and that filter brought us down to just a few.
"It was really in-depth and quite hands-on."
The final document summarized all of their findings in more than 100 points, spread across wide-ranging topics like the breadth and width of the feature set, how easy it is to use and integrate, the extensibility of the solution, the flexibility of the tech stack, the company's stability (how established it is, games that use the tech, what happens if it goes under), scalability (max number of CCU per project per region, likely bottlenecks in scaling, response to sudden load changes), and support (promised turnaround time, satisfaction, process).
The People Can Fly team even created a comparison matrix and dropped all of their existing games in to see what features would be supported and ran deep dives on cost for each game. Finally, the whole thing was color-coded from best (green) to worst (red) and weighted.
That narrowed the list to just a few companies, including Pragma. The vetting process happened to overlap with the studios building gameplay prototypes, so People Can Fly ended up taking all three finalists for a test drive.
The reasons
During the process, the team determined a number of critical features that were important for the company to achieve its goals with its in-development and future games. These features were among the reasons why People Can Fly selected Pragma.
Managed or Self-Managed
With a managed infrastructure, the backend engine company hosts and operates the backend for your game. With self-hosting, there is added complexity, but that comes with full control and freedom.
Pragma offers the flexibility of both options but most of their customers opt for managed.
If a studio decides it wants to use a managed infrastructure, then from day one the studio can deploy its backend to cloud-hosted shards which enables it to test and develop with distributed teams. As the studio’s needs for more shards grow, Pragma can stand up as many as it needs, and when it’s time for the studio backend to scale to millions of users, Pragma will help the studio scale up with load tests, code reviews and an experienced live ops team.
When it comes to using a self-hosted solution, Pragma offers the freedom and control to host Pragma where and how the studio sees fit. It can require a larger team of infrastructure and live ops engineers, but for companies that have existing infrastructure and dev-ops teams it can give them complete control.
"From the onset, there was one big question: Do we actually stick with a fully self-managed solution or go with a managed solution?" Roche said.
People Can Fly knew that a managed solution is typically highly proven to scale, but also knew that it would put the studios in a position where they wouldn't have a very high level of control if they needed customizations.
"There's pros and cons. We had to figure out if that was really the interest of the teams and the studios or the projects that were operating internally," he said.
People Can Fly liked Pragma’s take on a self-managed service, which gives studios complete control and customization of their platform while delivering massively scalable architecture, rich integration patterns, and source code access.
"We like being able to have that source code,” Roche said, being able to deploy with the skills and knowledge in-house using Pragma helps mitigate a really big risk for us."
Flexible, customizable matchmaking
Not all games use the same approach to multiplayer support and matchmaking. Even among the games in development by People Can Fly, the various games use different approaches to getting players in a game together and keeping them in sync, all ways supported by Pragma.
“There was no longer a concern about different server types because we can build a service within Pragma to actually solve that problem for us, integrating it more closely with the accounts and the inventory management that the service provides naturally,” Roche said.
Also, without Pragma, the studios were struggling to do real-world matchmaking tests and things like player data services.
“Being able to bring in Pragma’s services lets us complete very critical parts of our gameplay flows.”, Roche said.
On top of that, Roche mentioned the team was impressed with how modular the solutions were with Pragma.
“Pragma allowed us to build custom matchmaking strategies for a wide variety of games,” he said. “It’s also integrated with other platform services, eliminating duplication and various scalability risks.”
Integration
Once People Can Fly signed with Pragma, the rest of the internal teams began to ramp up with the integration. Every team had about two people that were helping, Roche said. “The process only took several weeks”, he added.
"Once it was done, it became almost plug-and-play, at least for that simple layer of connectivity and functionality with the other projects," he said.
Now, People Can Fly uses Pragma across multiple projects despite each game using various server approaches.
"We're making very good use of the features that Pragma provides and even the things that we aren't using today—like most of the features on our vetting list—we're going to evaluate it at some point as we develop the games," Roche said. "They've been very proactive. I feel like they gave us a bit of a red carpet rollout when we onboarded, and they were assisting us with the integration."
Looking back now at the entire roughly one year-long process of vetting a broad swath of backend engine companies and then testing a handful with in-development games, Roche says that approach was very valuable.
"If I were given the choice to do it again, I probably would do it more or less the same way," he said. “We gained a lot of insight that I think would have been difficult to gather just by reading product brochures."
Learn about People Can Fly on their website.
About the Author: Brian Crecente runs consulting agency Pad & Pixel, LLC. He founded Kotaku, co-founded Polygon, and was the video game editor for Rolling Stone and Variety. Before his time in video game journalism, Crecente was an award-winning newspaper reporter who covered crime and public safety for more than 15 years.
