Notionally, cloud gaming (or game streaming) is like playing your game on that really good machine your neighbor has, but plugging in a controller with a really long wire that reaches your room, and pointing a camera at the screen. Then all you need is a screen device to capture the camera stream. Given that we accept the cloud’s dominance with other computing loads, why not with the high power needed for modern gaming? And will similar issues impact the upcoming metaverse?
Google has been blamed for the failure of Stadia, which is fair enough — if you look at the Google Graveyard, you will see that they kill off the vast majority of their children with aplomb. A lot can be said about their lack of connection to the communities they should be working with, but that has always been the case. Project teams within Google know the score: if you can’t scale to the mothership’s recipe, you will be dropped without warning. Very much Death or Glory. I can only hope the game developers left in the lurch can dust themselves off and learn the appropriate lessons.
Having been involved in telecommunications as well as the games industry (as well as the telecommunications industry getting into the games industry), I’m a little more interested in why such a certainty as cloud gaming hasn’t come into its own yet.
What Is the Market for Cloud Gaming?
The games market has always been a mix of whim and possibility; there is no true linear logic in what the next big thing will be. The cost of using Stadia was high, but not expensive compared with (say) purchasing a new iPhone. Certainly, for a processor-hungry game like Cyberpunk 2077, it might be the only feasible way to play for many. Yet the black hole that Stadia fell into was that of “market fit”; they persuaded neither gamers nor enough developers that they had a strong commercial offering.
So will there be a market future for this type of thing?
There are already solid cloud gaming platforms, from Microsoft, Amazon, PlayStation and Nvidia. A quote from the magazine Polygon provides the best summary of how the remaining industry players see it:
“It is not a paradigm shift, it’s just an added boon.”
The idea of cloud gaming as a breakthrough product has probably now been exhausted (the previous attempt with OnLive failed in 2015). The concept of making a highly detailed game with little interaction beamed to your tablet is at first appealing, and this did spur on some entrants to the field. But soon the realization hit that you can’t create a great game design by starting with a delivery technique. This is just not what gamers are playing, nor what designers are making.
As the games need to be running and maintained centrally, the cloud customer sees ownership replaced with service. So Stadia was a content provider, much like Netflix or Spotify. This does cause some consternation but doesn’t appear to be a problem for most players if there is an appropriate upfront saving and a wide range of products. People do replay games or stay in one game for a long time, but these tend to be the games that need horsepower to run. However, the nagging feeling remains that games are not consumed in quite the same way as movies, books or music.
Smarter Separation of the Workload
From a technology perspective, Stadia was fine. The enemy of cloud gaming remains lag (caused by the round trip between the message that you pressed, the fire button going up to the server, and the response images of bullets spitting out coming all the way back) but this should slowly improve. The internet is not about doing smart things over the wire, it is about smart things happening at the edges, with the transmission remaining simple.
You could argue that once your images have been created, just sending them to the target device is a reasonable — if heavy — use for the internet. But this is the other difference with gaming; unlike a movie or a book, the human is continually interacting in an active way. When my PC overheats I know it will lose frames (i.e. things on the screen will jump) but at least I can control that. But small inconsistencies in any visual, aural or haptic feedback from a remote system that you are relying on will quite literally get on your nerves.
While I am a big advocate of Network Neutrality, the fact that no internet provider can give me a line whose network speed is truly guaranteed not to let me down is a big a part of the problem. More to the point, a cloud gaming company can’t compel the internet provider not to do work on the line at strange times of night, when many gamers are active. Even if the network provider and the cloud gaming provider are the same company, they may not get exclusive access to the line at all points. Consumers do not want to be in the position where no one will take the blame for an outcome, with each service blaming the other. We’ve all been there and nobody likes it.
Improvements will likely come by smarter separation of the workload. Why, for example, would I want a static menu generated upstream?
On the other hand, streaming a cutscene that I only glance at once is a tremendously sensible thing to have on the cloud. In fact, knowing a dull cutscene many MB in size has to be stored on your own device is kind of annoying.
Asking separate services to render parts of the game means that technically, multiple services could offer to help. While this sounds like some type of microservices hellscape, it stands to reason that (for example) studios that dedicate themselves to perfecting cutscenes can deliver these as needed. Many games add a physics engine, but all sorts of dynamic effects could be outsourced. Multiplayer platforms have been tentatively offered from companies like Improbable, Polystream and Amazon Lumberyard (which is now Open 3D Engine). We’ve seen attempts to integrate separate voice chat or achievement systems from other platforms; most studios know they don’t have to do everything anymore.
Impacts on Metaverse
So how does this impact the metaverse? While no one knows exactly what the metaverse is yet, it will at least be a 3D visual experience, in a persistent world with a large number of people experiencing it in real time. The likely high computing cost of this type of platform, added to the hardware addition of a headset already being a necessary part of AR / VR, will limit the delivery options that will work for the audience size that Meta (for one) is clearly aiming for.
Streaming is very likely to be the only way to make the metaverse accessible to a mass market.
Unfortunately, most of the negative points about cloud gaming also apply to streamed virtual persistent worlds. People will complain when their cozy world is sold to a company they don’t like. They will wonder who is to blame when performance degrades. And performance problems will be visible to vast numbers of people at similar times, allowing them to band together to complain. Historically, interactions between avatars in (for example) Second Life have been relatively sedate, but it would be unwise to limit future worlds with this perspective.
Compared to books and films, video gaming is a young entertainment form. Patterns of behavior are still being born. We knew that the Kindle was good. We suspected that 3D films were unnecessary. While it is a bit harder to focus clearly on the exact value of cloud gaming, the evidence points to slow but steady adoption as problems are smoothed out across the board, as opposed to sudden innovation from one point.