There are people who play games and there are people who watch games. While it is debatable whether Video Gaming is a sport, it is a fact that Video Gaming rivals traditional media as a form of entertainment, whether offline or online (i.e. available over the internet). As Netflix stated in its shareholder report that “We compete with (and lose to) Fortnite more than HBO”.
As online gaming (mobile, PC, console, etc.) continues to exponentially increase, the network performance (not just availability) is critical for the superior game experience and in particular to action games such as FPS (first-person shooters) and e-sports in multi-player mode.
Arguably, network performance may be more critical for cloud gaming such as Google Stadia, since the gaming experience solely relies on the cloud (for almost all of the processing) over the network. The network performance (mainly, bandwidth) is also somewhat critical while broadcasting one’s game in real-time on any of popular social platforms e.g. twitch, youtube etc.
for others to watch (note that online game watching is the second most popular viewing with~100 million viewers, more popular than MLB, NBA or NHL.
This paper covers the challenges in getting an excellent online game experience for action games, not only from the game developer and players point of view, but also from the network point of view. The paper does not focus much on game streaming (e.g. Twitch) or Cloud Gaming (e.g. Stadia).
As we discuss in the paper, the online game experience wrt the network is multi-domain by nature – home network domain, service provider network domain, Internet domain and the data center network domain (where the game servers are located) are all important parts, each one with its own set of technical and non-technical issues. Of course, there are other challenges in ensuring superior game experience outside the network, such as the game engine, match making, rendering rate of a graphics card etc., but these are not the focus of this paper.
We see a trend of service providers marketing “low latency” as a differentiator that is essential for gameplay. They equate “Lag” to “Latency”. It intuitively makes sense. Isn’t multi-player gaming similar to a dual where the fastest to draw is likely to win? As we explore in this paper, the answer is more complex and depends on the type of lag compensation algorithms used by the game server, and what the gamer defines as “Lag” is not exactly what a network expert defines as “Latency”.
The above has been the trailer to our paper, now let the game begin!