In 2001, game designer Raph Koster and producer Rich Vogel presented a talk at GDC about persistent online worlds and what exactly it would take to create one. Almost two decades later, it’s quite interesting to revisit the challenges that they outlined and explore some of the barriers that are still standing (in addition to new ones that have since popped up).
At the time, Ultima Online had introduced the concept of MMOs as we know them, and the incredibly successful Everquest had really pushed the genre to the forefront of the industry, but MMORPGs were still quite young and raw. Interestingly, at their talk in 2001, both Koster and Vogel identified the average lifespan of these games as 5 years, which, in the case of Everquest, was exactly right (it remained the category leader until 2004, when World of Warcraft was released). They similarly had the foresight to identify the genre as a service (rather than product), and outline the ambition to support thousands of players instead of a few.
They specifically identified the issue of scaling as a critical consideration, accounting for service components such as player profiles, achievement ladders, and small group infrastructures, as well as the ongoing addition of in-game content. Live development is one of the biggest changes we’ve seen in the last couple of years and is redefining the way studios and developers work (we’ve gone into this in more detail in previous blog posts around how AAA studios are changing and Ubisoft’s live games and live development). Certainly, from a content creation perspective, there have been significant strides in the efficacy of the front-end tools available to studios, developers, and artists. The adoption of commonly known game engines like UE4, for example, has broadened the industry’s talent pool, while more specific applications, such as Unity’s recently announced Kinematic software, have made the creation of art and animation assets much speedier.
Beyond this, Koster and Vogel also outlined server loads, security, backend support, and database management as significant challenges - all of which remain so to this day. The problem is that, over time, the size and complexity of these game worlds has gotten significantly large, compounding the difficulty of these issues. In 2001, the pair accurately predicted that the size of persistent online game worlds would continue to grow from the approximately 16x16km maximum size of that time, but they probably didn’t imagine anything at the scale of Minecraft’s game world.
EVE Online, home of perhaps the most famous persistent online world, has run for the past 15 years, built on a proprietary infrastructure and technical workarounds to maintain the game experience. It’s a clear example of the ambition of a connected game being held back by the constraints of modern technology. The game clevelry makes use of a grid system that only activates grids of high-definition space that only pop up around the immediate proximity of a player - however, this is largely achieved through quite complex and fragile server manipulation. When certain server nodes started running out of memory in the mid-2000s, filling up with legitimate user data and crashing, CCP (developers of EVE Online) controversially introduced. player limits to star systems under heavy load.
Persistent online worlds are growing and evolving, but to date the gaming industry hasn’t found the ideal technical solution to support the kinds of experiences that players want to play and developers want to build.
HadeanOS is a cloud-first operating system that has been engineered and optimized for performance across massively distributed computing infrastructures. HadeanOS natively understands the dynamic scale and real-time demands of modern applications in the cloud and removes the need for complex operations and engineering.