-- RobertLarese - 9 May 2024

An application of the "cloud" to free computer gaming

When you hear the phrase "computer gaming," bulky consoles come to mind, where players run games on their local machines. Then there is another set of third parties that provide access to group-playing, so called "going online" by connecting local machines together over the internet. Bill Gates is happy to sell you the local machine, and for an allegedly paltry ten dollars a month, provide the connectivity to other local machines. But this type of gaming is comprised of at minimum personal hardware running the game locally, should the "gamer" forego multiplayer.

Enter cloud gaming.

Imagine an end user, the player, a pipe relaying the user's inputs to a remote machine, could be a server farm or even another's personal computer, where the game is run, and a pipe that transfer outputs back to the user, over an internet connection strong enough to handle the game's rendering requirements. It is like Netflix: "Cloud gaming . . . renders an interactive gaming application remotely in the cloud and streams the scenes as a video sequence back to the player over the Internet." There is something freeing about playing a game with the mouse and keyboard of a computer smaller than Hyacinth Bucket's ivory slimline telephone. Cloud gaming's appeal is instant: It offers users high-powered computing without the computing limits of personal hardware. None of this is new.

Cloud gaming is just another application of the "thin client" principle. All this is is the ability to run programs on one computer but engage with them on another computer. Citrix has been doing this with respect to business computing since the 1990s. And anybody familiar with turn of the century corporate America can recall the loaner laptops companies kept on hand for the rare times employees had to work remote, enabling employees to pull up their work desktop virtually. Running Shogun: Total War virtually may be more demanding than MS Excel, but it is not new.

But it can be free.

Researchers at the University of Michigan in 2020 developed a DIY cloud gaming implementation, leveraging Moonlight and Google Cloud Platform (GCP), that delivered to users computer gaming at "approximately 60¢/hour when in use . . . (after) approximately 500 hours" of free cloud gaming for the first three months of use. If somebody were to offer you free postage if they could read your letters first, you would probably say no. GCP is exactly that arrangement.

Were you to investigate academic treatments of privacy concerns vis a vis clouding gaming, or even cloud computing generally, password security and data breaches would occupy your field. Providers muster their "password strength," power to ensure your data is safely in their hands for their use. Cloud gamers may expect the same passionate concerns from their providers, too. In many respects, Netflix may stand in for cloud gaming companies. "Cloud gaming companies collect and store a large amount of personal data from users, including their gaming preferences, purchase history, and personal information." Consider again, the Netflix analog. When users had to order DVDs, Netflix's development of user preferences required at least some physical interaction; now, they even know what movies you mull over but ultimately pass on. If you thought this was not a disaster already, Netflix's nascent expansion into cloud gaming might make you think contrariwise.

Recall the researchers at University of Michigan who built a free cloud gaming implementation that was "applauded internationally by (several thousand) cloud gaming enthusiasts from Latin America to Singapore." You can teach people to do something for themselves over the internet. The cloud gaming industry might be primed for a grassroots implementation of a more private gaming deployment.

What would that look like?

Amazon offers a cloud computing subscription where the user can set up a dedicated server to host games on. Parsec, who offers a similar remote desktop product as Citrix, offers a service that allows "a game developer . . . (to) run multiple versions of their game on one virtual machine and stream that to thin clients." The proposal offered here is to do exactly that except hosting on a private machine, not a rented server that is being monitored. The infrastructure needed to accomplish this is not prohibitively expensive, nor is the architecture complex.

Gaming servers are not unattainably expensive to set up. The scale is in the hundreds to few thousand. And as stated earlier the relevant virtual network computing implementations are anything but new. Developers have already proposed solutions. But most importantly what this paper proposes is not confined to niche or knockoff gaming titles.

There are several lists of top games that may be launched on a private, secure server. These include Minecraft, Rust, Factorio, Counter Strike: Global Offensive, and others. Other open source games could be self-hosted on a private server, too.

Cloud server companies call this solution impractical and "associated with additional acquisition and electricity costs" with respect to the servers themselves. They will sooner rent you one of their virtual machines. Others might say, "(u)nless you're a Netflix, Facebook, Amazon, et al, with dedicated product teams . . . required to stand up your own testing platform," it is not worth it. Of course, gaming companies are not only in the business of game development. They operate servers of their own to host games on, too, to support their data collection ambitions.

But this thinking misses the mark entirely. It is not even that important that everybody who likes to enjoy computer games with other users in multiplayer do so this way. That Instagram has over two billion users, does not mean that an open source, federated social media like Mastodon should roll over. It now has ten million users. Two people gaming together this way would be two freer people.