Online gaming has really come a long way since the acreation of NIMROD, one of the first custom-built computer games introduced in 1951. Gaming has evolved from single-player arcade games, to multiplayer consoles, to mobile gaming where players can take to the streets with their smartphones to seek out elusive augmented reality (AR) characters like Pokémon. Over time, online gaming has increased the number of interactive players worldwide, turning it into an ESPN-worthy spectator’s sport, estimated to reach 454 million viewers globally in 2019 according to eSports Trade Association.
Game developers continue to take advantage of a number of technology trends that will further drive innovation and improve playability. For example, more immersive gameplay through virtual reality (VR) and AR will support photorealistic graphics that are almost indistinguishable from real-life. These VR/AR gaming platforms will also open the way for players to connect and socially interact with each other in virtual spaces. 5G mobile networks will enable faster game download speeds, low latency and more simultaneous connections for grander multiplayer experiences.
Cloud gaming makes the next evolutionary leap by streaming games from any device with an internet connection (including smart watches and mixed reality glasses), just like music and video. Cloud gaming platforms can provide much needed compute and storage resources for open source gaming, which enables game developers to modify existing games and create new variants of games. This will remove the barriers to entry for smaller, niche game developers.
These innovations in gaming all have one critical dependency that could be a gaming experience killer: Latency.
Cloud gaming is already a crowded field with a common goal
Gaming as a service is nothing new. Sony PlayStation, LiquidSky, Vortex, NVIDIA GEForce, Parsec and Electronic Arts offer traditional and contemporary gaming experiences as a cloud-based subscription service. Hyperscale cloud providers, Microsoft and Google will also soon be entering the cloud gaming market this Fall with their Microsoft Xbox One “Project xCloud” and Google Stadia “Project Stream” streaming game services.
Regardless of the types of gaming capabilities and pricing models these providers offer, they all can agree on one thing…ultra-low latency is vital to the gaming experience. When Google pulled the covers back on its Stadia and Project Steam gaming as a service platform, at the Game Developers Conference last March, it spent much of its presentation talking about the horrors of latency on user playability. The company cited that everything about gaming has some amount of latency:
- HDMI Transfer: 16 – 33 milliseconds (ms) latency
- Monitor Refresh Rate: 4 – 16 ms
- USB Polls for Peripherals: approx. 8 ms
The chart below maps out how latency impacts player actions in online games:
But the greatest, most unpredictable impact on gaming performance is latency from either network congestion, lost packets, lack of available bandwidth or jitter (the time delay between sending packets in sequence). These issues are generally caused by a centralized gaming platform, where content must traverse multiple router hops and/or peering points across a long-haul network. These network infrastructures are not sufficient for real-time cloud gaming, where any buffering of packets or caching of content disrupts the constant stream of packets to its users, delivering an inconsistent gaming experience. What’s required is bringing cloud gaming closer to the user wherever they are and supporting any gaming device in the same consistent manner. This means literally taking cloud gaming platforms to the edge.
Dynamic cloud gaming needs a dynamic network
Cloud gaming companies know that their underlying networks must automatically sense what is happening in the network to scale and react to changing latency conditions and bandwidth requirements. That is the only way they can deliver the right latency and throughput to maintain consistent user performance and playability. This means gaming companies offering cloud services must have multiple distributed “availability zones” proximate to high concentrations of gamers, anywhere in the world.
Another critical dependency for all of these games is access to a peering infrastructure. Given that the large majority of these games are played over the internet, then the fastest and easiest way to get localized access to the largest population of players is to participate in lots of internet peering exchanges. Internet peering came about because no one network goes everywhere. A peering exchange enables multiple internet service providers (ISPs) to access a room full of different vendors’ routers. For example, sometimes for a player to get to a cloud gaming system that’s in a certain location, then Verizon and AT&T need to peer with one another to get them there. For many ISPs, that place is an International Business Exchange™ (IBX®) data center. Inside these interconnection and colocation facilities, we operate what we call a “peering exchange,” which businesses connect with to transfer traffic between one network and another.
The gaming companies, the cloud companies, the networking companies all participate in these peering exchanges, which are geographically distributed around the world. Using North America as an example, we operate IP peering exchanges in Toronto, New York, Washington, Atlanta, Miami, Dallas, Chicago, the San Francisco Bay area, Seattle and Los Angeles. If you were deployed in all those different peering exchanges and drew circles around each of those IBX facilities, you can realize less than 20-millisecond latency between the vast majority of people in the U.S. and those exchanges, which is more than acceptable for most big games. Those gaming companies that are in four or more of these peering exchanges across the U.S. are going to have better performance than a gaming company in just one or two of them.
Leveling up in the cloud
Gaming companies have long seen the value of the cloud for auto-scaling the computing power required to stream games to multiple players. For example, you can leverage a bunch of virtual machines in the cloud and scale up compute capacity pretty quickly. Years ago, I was at an event where Microsoft was launching one of their Xbox multiplayer games. They demonstrated that within five minutes of launching the game, the system could spin up 80,000 cores in their Azure cloud to deal with the increasing load.
However, in native cloud gaming, the idea is to get rid of the console, which in itself is expensive and, as we mentioned above, has its own inherent latency issues. By transforming online gaming into a pure cloud play, developers along with hyperscale cloud providers such as Microsoft and Google, can harness their presence in distributed peering exchanges.
Take the cloud-based gaming company Blade and its gaming platform Shadow for example. This Paris-based business had its eye on global expansion and multiplying its 20,000 users by 10x. It also had a goal to improve the performance for its online gamers by providing <16ms latency worldwide. Blade based on its global interconnection and data center platform to help it achieve the double objective of increased growth and performance. Blade had a strong international presence, especially in the United States, and a scalable and secure infrastructure. But it was the concentration of numerous telecoms operators and network platforms, including those that could be accessed , that helped them achieve the lowest latency possible in the online gaming industry.
A number of our other gaming customers, such as Pearl Abyss, Tencent and Zynga, have recognized the importance of an internet peering infrastructure, in addition to a highly concentrated ecosystem of cloud providers on Platform .
As online gaming takes to the cloud, game developers, hyperscale cloud providers and players are all winning by harnessing the power of scalable high performance, low-latency interconnection at the edge.