It’s the ultimate vision for digital entertainment: generate high quality content in a massive data centre, then stream it instantly to any device, anywhere.

If Netflix has popularised this idea for movies and television shows, why can’t Google, with one of the world’s most powerful internet platforms, do the same for video games?

That question hangs over Google’s forthcoming “cloud” gaming service, Stadia, which was announced earlier this week. The gameplay all happens on servers in Google’s data centres, rather than a player’s own hardware, and images are streamed to any of a user’s screens. On PCs, players can use game controllers they already own. For other screens, they need to buy a new one made by Google.

It all sounds deceptively simple. Other companies, including Microsoft, Sony and Electronic Arts, are working on similar ideas. But cloud gaming throws up several tough technical challenges. Whether Google can crack the problem by the time it launches Stadia later this year will be a good indication of whether gaming is ready yet to head to the cloud.

There are three main challenges. The first is to create visuals that match or exceed game consoles and high-end gaming PCs. Google has set the bar at producing images in ultra high-definition — also known as 4K, which is ahead of today’s consoles — and at a fast enough frame rate to make movements appear smooth (Google is matching the 60 frames per second of the latest Xbox and PlayStation consoles).

In this department, Google has home field advantage. Its data centres, built to support mass consumer services such as Google Maps and YouTube, are among the world’s most powerful digital content factories.

To render images for its Stadia games service, Google has built new server “blades” — thin machines that can be slotted into racks in its data centres. The Stadia machines are embedded with custom-designed graphical processing units produced by AMD. In terms of raw computing power — defined by the number of operations a chip can produce in a second — that means more than 10 teraflops. The current console champion, Microsoft’s Xbox One X, manages 6 teraflops, though the next generation of consoles, which could come as early as next year, is expected to go much higher.

Running games in a data centre opens up other possibilities. Google said it would be able to call on multiple graphics chips to render the most complex images, like a waterfall — something beyond the capability of a standalone console. It also said it would be able to let many more people take part in the same multiplayer game at the same time, creating the possibility of “battle royale” games such as Fortnite, in which thousands can compete to see who is the last player standing.

The second big technical challenge for cloud gaming is to stream the images to a user’s screen over a fast enough network to avoid degrading the experience.

Google has developed its own compression algorithms, squeezing the size of the data files to be able to transmit more information over the same amount of bandwidth. It also has its own communications “backbone” — more than 400,000 miles of fibre-optic cable that link its data centres around the world. This means it does not have to rely on the public internet, which cannot deliver high guaranteed levels of service.

So far, so good. But it is when the digital bits travel beyond the reach of Google’s network that things get more problematic. Players will be dependent on the internet connection they get from their internet service provider. Google said it recommends broadband speeds of at least 25 megabits per second to deliver high-definition images (known as 1080p), though it also claims it can get to 4K — with four times as many pixels per image — with only 30 megabits.

Some potential customers won’t have enough bandwidth. It also limits playing games on mobile devices, unless they are connected over Wi-Fi to wired networks. Future upgrades to mobile networks will help, but they may be years away, and Google itself made no mention of the potential of 5G networks.

“5G is being described as a ‘saviour technology’ — but at least for the next five years, we’re not going to see mass deployment,” said George Jijiashvili, an analyst Ovum.

The third technological challenge for Google will be to eliminate the latency, or lag, that can slow the response time when a player presses a button on the game controller. A few microseconds between the action and seeing a response on the screen can make all the difference.

One answer is that Google claims more than 7,500 nodes, or endpoints, to its computing network, increasing the chance that players will be close enough to see a fast response time. Even that, however, may not be enough: many users outside large cities could find themselves many miles from a Google node. For people in countries such as Australia, latency is a notorious problem, Mr Jijiashvili said.

Another way Google hopes to reduce latency is to cut out the number of steps a player’s command has to take before reaching its data centres. Its service does not rely on a streaming box to carry signals to a TV screen. Instead, its new game controller is connected through Wi-Fi and communicates directly with Google’s data centre.

For all its efforts, though, Google will still be at the mercy of other network operators. “Because these games require high speed and low latency, I have a feeling that the experience the end users will get will be out of Google’s control,” Mr Jijiashvili said.

richard.waters@ft.com



Source link

LEAVE A REPLY

Please enter your comment!
Please enter your name here