In the initial reviews of Google's Stadia platform, many cite latency as a significant issue for the quality of the experience, especially over a wireless signal. This was the downfall of many previous game streaming services, including one of the first live game streaming services, OnLive. Even the ones currently in existence don't do very well with latency-sensitive games like FPS. You can get away with some of the latency with casual fighting games, sports games and racing games, but the reality is that fundamentally, gaming is much more about latency than it is about bandwidth.
The issue is that latency is introduced in virtually every step of the process—from the wireless input of the user's controller into the ‘console,' all the way to the latency it takes to render a frame on the GPU inside of the Google Datacenter. The networks of today still have too much latency to stream games without noticing the problem. There are techniques to mitigate latency, but nothing Google does seems to be effective. Both 4G cellular and broadband wireline internet services still carry about 50-60ms of latency end to end, and that doesn't even account for the controller's input latency. All of this said, what's to be done?
5G to save the day
To lower the latency, the streaming infrastructure will need to get closer and closer to the edge of the network and closer to the user. A component of this is edge computing, or physically bringing the processing capability closer to the user near the edge of the network. The arrival of standalone 5G networks could provide that lower latency needed for gaming titles like FPS. That's not the kind of 5G networks that we have today, which are non-standalone and heavily leverage 4G infrastructure to power 5G radios to get 5G to consumers faster. Standalone networks already show latencies below 30ms, but I believe we need to go below 10ms to really have a solid experience. Another opportunity to lower latency inside the 5G network may lie in network slicing, in which a specific piece of the network can be provisioned and optimized for game streaming services. This in theory could deliver exactly enough bandwidth and latency for a good user experience.
Another thing that's holding back game streaming is the amount of bandwidth that it takes to stream content at HD and 4K resolutions. Gaming for hours on end could result in consumers blowing through their data caps, which unfortunately exist for both home broadband and cellular connections. This should theoretically be less of a problem with 5G, as data caps get bigger and some operators move towards unlimited. Some operators have even launched their own game streaming services in an attempt to get around the bandwidth issue while simultaneously increasing their ARPU.
Others will enter the market, like Microsoft and its Project xCloud. People's experiences seem to be better with xCloud, but it remains to be seen if it's a worthy substitute for an actual Xbox console. Microsoft's messaging indicates we're not quite there yet.