Cloud game streaming (like VDI for gaming) is coming—Time to re-apply all the lessons we learned

Will it work and bring a form of remote computing to the masses? Who knows! But we can anticipate some of the questions that will come up.

According to various sources, a big wave of cloud game streaming offerings is on its way. For those of us in the desktop virtualization space, we know what this means—games running in datacenters, with gamers playing them remotely via a two-way protocol.

Cloud game streaming means that a huge industry vertical and its customers are going to be learning the lessons that the desktop virtualization folks have been known for years. This is clearly something to keep an eye on, and that I’m sure could translate into some big opportunities for desktop virtualization vendors and professionals.

What streaming gaming services are out there?

You might remember OnLive, the non-Windows licensing compliant game streaming startup we covered in 2012.  OnLive was founded in 2009, but the original incarnation was out of business by the end of 2012, having never gained much traction.

But, now it’s 2019. Remote graphics technology has come a long way, and earlier this month we saw a spate of stories involving tech giants. Verizon is quietly testing an offering called Verizon Gaming; Amazon is supposedly working on something; Microsoft is working on an offering called Project xCloud; and NVIDIA GeForce Now is in beta. Google was working on Project Stream, but now the beta is over with no word of a release.

This is on top of a couple of other fledgling services, including PlayStation Now, Blade Shadow, some games on Steam, and a few others.

Classic VDI arguments

Cloud game streaming clearly has a lot advantages that are similar to VDI and other forms of desktop virtualization:

  • Some games are huge—like 50GB for many AAA titles—so playing them remotely can be much faster than downloading them and running them locally. Many games also have Day 1 patches that are another couple of GB, but a cloud gaming service could just manage that, instead of the customer having to wait for it to install.
  • The game platform can be different from the client platform.
  • The cost of playing moves from capital expenditures—a $500 console or $1,500 PC and $50 to $60 per title—to a monthly subscription model.
  • Subscription models, in turn, give players easier access to more games, and developers easier access to more customers.

And there are the disadvantages:

  • Cloud gaming requires players to be online, and the connection has to be decent (which is frankly hard to find in some regions).
  • Gaming is very sensitive to latency and temporary connection issues.
  • Plenty of gamers still prefer to own physical media, and PCs and consoles are entrenched in the market.
  • A remote experience will always lag a few years behind a state-of-the-art native experience.
  • Apparently some services charge by the hour, so the cost could get out of hand, if you’re not careful.
  • For services that charge a monthly rate, high-end VM instances could affect the cost model real fast. (Just think of the blowback they’ll get if they have to throttle users or charge overage fees.) And remember, as the years go by and computing gets cheaper, game software is just going to keep demanding more resources—there’s no outrunning this problem.

More issues to think about

Beyond the standard pros and cons I just outlined, there are a million other VDI and desktop virtualization conversations that apply directly to cloud game streaming.

There are a lot of different options for clients: Web, native clients, PC, mobile, thin clients (think lightweight game console). All these will affect the experience. Does the client have a GPU? A specialized SoC? The appropriate input devices for the game title?

It will be interesting to learn what type of infrastructure these platforms use—I’m sure they’re not just installing a copy of Horizon or XenApp and calling it a day. And will Windows games be properly licensed?

How about the remote display protocol? It will have to be tuned completely different for game play versus a VDI desktop or even a remote CAD workstation. (The protocol wars may happen again in a new form!) 5G networks are on their way, so naturally there are lot of predictions that they will help.

Customers might naturally adjust their expectations, too. Some players can get used to higher latency or change their style of play (strategy within a game, what type of character they choose, etc.). If you want to play online with a friend on another continent, you know that you’re going to have to put up with some latency issues.

And finally, companies have to be getting into it for the right reason. I mentioned that cloud game streaming lends itself to subscription models, but of course you don’t need that technology to do that. There are already subscription services that use locally installed games.

Final thoughts

Thinking about all of this, I couldn’t begin to predict whether cloud game streaming will be commercially successful. Kyle, our resident gamer, likes the idea, but it’s more out of curiosity—it’s not something he’s been waiting for forever or will likely try anytime soon.

Either way, there are a lot of opportunities for the desktop virtualization and remote graphics community to be part of the game streaming conversation.

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.