Google Teaser? #Stadia

Another console to include?

https://stadia.dev/blog/unity-support-for-stadia:heres-what-you-need-to-know/

 

This is 100% going to be a streaming device, and knowing Google, it'll only be available in America.

So non-starter, really.

 

This appears to be a game-changer... for players with good broadband, and I should imagine publishing houses with plenty of cash.

 

All about them streaming data-centers, son. Anyone who thought this was going to be dedicated hardware was kidding themselves.

It is nice to see that they are actually putting out a specific controller. I don't much care for the D-Pad on that thing, but the idea of it connecting via WiFi is actually pretty great. Given the approach they are taking, it makes sense not to bother connecting the controller to a specific device.

The real advantage of this whole thing is how easy it will be to get people into it. Low barrier to entry is going to make it easier to expose it to a broad audience.

Ooooo, I just got to the part of the presentation where they show off features for local multiplayer performance benefits, that's pretty cool. I could do some nifty stuff with that.

 

Can one use unity or unreal to make games through stadia ?

 

Yes. They are pushing a streaming platform, not a specific engine. Pretty much any existing game engine will likely be able to be deployed on their platform. You would just have to use whatever API they provide to tie their services into your game. Both Unity and Unreal have hooks for incorporating external libraries, so there won't be any limitations there.

 

Their statement was that they partnered with Unreal Engine and Unity to enable developers to make their game run on Stadia (which is a Linux platform).

 

Yes if I go to stadia.com looks like is not ready or is not available in my country.

Unity Support for Stadia: Here’s what you need to know:
https://stadia.dev/blog/unity-support-for-stadia:heres-what-you-need-to-know/ <-------<<<

 

I'm hyped. You guys?

 

Stadia link system, as was presented, will be very engaging and I hope not intrusive.

 

I'm a little bit hyped. The lower barrier to entry is going to make it a lot easier to get people to try games on this platform. One of the big issues I foresee is input disparity, but it's not like that isn't already an issue for any PC platform. It would be mitigated somewhat by their standardized controller, but not entirely stamped out.

It's going to be a bit of a beast where bandwidth is concerned. Anyone with data caps is going to be harder to reach. That is one of the bigger factors that is going to throttle progress on a platform like this. Thankfully, the demands of video streaming have already started to push the demand for better on-line infrastructure. Everyone wants a fat pipe these days.

As far as the development potential, I'm actually pretty excited. There are a lot of designs that this approach to gaming would make possible. For developers, there are very few downsides. It would effectively secure you against piracy, eliminate the need for updating clients or pushing downloadable content, and eliminate hacking, cheating, and bots in on-line games. There are a lot of upsides.

 

I'm definitely curious about how the latency feels when in a fast paced action game, but I don't play too many of those.

There is a scaling problem on the network side, ISPs would probably start capping the traffic if it eats up too much bandwidth. Just like Netflix there will be finger pointing as to who has to pay for upgrading the network infrastructure and they'll drag net neutrality into it and oh fun times...(assuming it got popular)

 

Yea I wonder how fast your internet connection has to be in order to get 4k at 60 fps?

 

I wonder what it will mean for indie gaming, and for games aimed at niche markets that don't conform to Goolgle's / societal norms.

I noticed from the wall of "partners" behind the speaker that Microsoft, Nintendo, and Sony were not present; as they make consoles that appear of slightly less value in a brave new world. Why would you need the next console upgrade if gaming doesn't involve your hardware any more?

Internet speed is another thing. I wonder what the minimum speed needed will be. There is still a good number of people with <5Mb connections. 2Mb is the max in my area until you hit the outskirts of the nearest town.

 

You think it's a coincidence that Google is working on their own fiber network? Create the demand, and then provide a solution to the demand. Those guys are looking to make money on both sides of that equation.

 

You mean that same fiber network where they just went "actually, this is hard" and then pulled out of an entire city?

 

I would not be surprised to see Google Fiber sold off to some of the regional providers one of these days, it is pretty much at a standstill for a long time now. The last mile is a very expensive proposition as they found out.

Just check wikipedia "In October 2016, all expansion plans were put on hold and some jobs were cut.[11] Google Fiber will continue to provide service in the cities where it is already installed."

 

This might be the push Unity needs to get IL2CPP support for Linux going. The link from Alan above says Stadia will be IL2CPP, but I can't imagine there would be much reason not to bring it to Linux stand alone at that point as well.

 

Why wouldn't the compiler planned for Stadia not work for Linux standalone? Is it a Stadia specific compiler from Google?

 

The thing is, as far as I can tell it is dedicated hardware. Sounds very similar to a console to me, it's just in a datacenter instead of in your house.

A few weeks ago when someone was talking about MS doing this I remarked that from a trial run perspective it'd make a lot of sense to tweak the Xbox One OS and put them in racks in data centers. This actually sounds a lot like that, except that Google have had to make stuff rather than using existing off-the-shelf tech.

 

It doesn't require a box at your house, and it can be run on all your devices. Also, it doesn't take any downloads, just streaming. How's that not hardware?

 

"The future of gaming is not a box."

Not responding to previous poster, I just think the slogan is hilarious.

 

Unity Support for Stadia: Here’s what you need to know
Unity Blog:
https://blogs.unity3d.com/2019/03/19/unity-support-for-stadia-heres-what-you-need-to-know/

 

I guess I'm just an old skeptic at heart, but am I the only one not on board with this?

 

Same, but then I'm not excited by anything that requires better a better connection to the Internet than is in my area.

 

I guess most hard core PC players don't care about Strada, and some console players category prefer to own a console.

 

The entire idea is a mess. No matter what changes they make to the back end, it's still the same old attempt at game streaming that people have been trying to push since 2009.

 

I'm in a poor country where Stadia is not supported but we use Netflix and I was able to download Unity Megacity at 1,2Mb per second. So I mean in practice can work but they need a lot of gpu.

 

What if it's not what if it uses AI to govern the stream and console hardware level to optimise server vs client load?

Or just takes advantage of cutting edge GPU/CPU/DRAM SOC hardware to allow cheap server side rendering of games.

 

https://twitter.com/krisgraft/status/1108121139815964673

Guess again.

Also the "console hardware" is a Chromecast device.

 

Yes, it's at least as much as a game changer as these services:

OnLive
Geforce Now
Shadow
LiquidSky
GameFly Streaming

Which is not at all.

 

It's a horrible idea at this point in time for a lot of games. Why cloud providers love it should be obvious. It's specifically bad for developers making high end games.

As a developer you now have to essentially pay for the cpu/gpu resources that otherwise would be payed for by the player. And on PC games that leaves players paying a premium to NOT use the perfectly good PC they have. Because one way or another that cost is going to get passed down.

How in the world does that make sense for anyone except the cloud providers?

That doesn't even get into the bandwidth costs.

And even google doesn't have a way to reduce the cost per unit of processing power by all that much. Most of the optimization they do is in how to cheaply power and cool it. Plus this is likely all bare metal, I don't think there is any cpu virtualization good enough for games. So that's not going to be cheap because virtualization is how they make money on processing time.

 

And knowing google, it will be abandoned in 24mths.

 

https://twitter.com/BitHunterOK/status/1108101326540550146

 

What are you referring to by CPU virtualization? Are you referring to hardware accelerated virtualization where you assign a number of cores to a virtual machine? Or are you referring to cores being shared by multiple virtual machines? Because the former is fully achievable with only a very minor (typically around three percent) performance loss.

 

Sure, the bandwidth is there in many cases. But what about latency?

 

That difference is just a configuration in the host OS. While you can get really good performance with virtualization it requires direct access modes, which cloud providers almost never use even when they dedicate cores. Because that would be a massive security issue. The modes that isolate guest OS's enough to be used in a shared environment, don't perform nearly as well.

 

If they're using custom, dedicated hardware I don't see why this is a concern? They're short on details, but to me this sounds like slot-in mass-produced nodes which each run one instance of one game.

Do you think that's really the market, here? If someone is already interested in buying / building a high end PC, what interest would a game streaming service have to them?

The high end PC crowd are also the people most sensitive to issues like latency. I don't think latency will be the showstopper which some people claim, but it will be of concern for some games and some people, and those most sensitive to it will still get the best experience by playing locally rather than remotely.

I predict that exclusive games will be designed to look gorgeous while elegantly handling latency.

 

about latency? server-side (here is around 100ms since 98).
It can cost like 9 usd per month like Netflix: 5~10 users pay for the chip in one year.

 

lmao, it's going to be using 7 gigs an hour for 4k. You might as well just download the damn games normally.

 

It's a game changer if it works. By game changer - no, players who have a console or buy high end pcs probably won't bother. But the vision is of more than that. Mobile gamers being able to access games which will drain less of their devices battery, and switch to the tv when they get off the train at home. Watching someone play a game on stream and being able to click a button to queue up and join as player 2. Throwing more boxes at your game to up the fidelity rather than making concessions because you have to run on an iphone 8. And damn if the prospect of a developer's next game going "eh, we're going to run the game on 4 gpu instances because we want to up the fidelity" without any players needing to upgrade their hardware wasn't exciting.

The details and execution sure, are big unknowns at this point. Having to run your game as a service is a pretty big downside, especially in terms of what unity is going for in being able to make a creative platform for anyone to express themselves with. No way you can distribute your game at no cost here, where the download model lets a solo or small group spend time instead of money to develop and launch. And we've seen the tech work very well with the Creed demo last year, but what will the reach, internet speed requirements, capacity, latency etc all really look like.

And cost cost cost. How much do we have to pay for these instances. I mean, ec2 is pretty cheap, but even running a simple website with decent traffic can add up. We are giving each user 1 or more gpu time for their instance? As they play for 2, 3, 4, 5 hour sessions? And how much are they paying? I would never buy a streaming-only assassin's creed title for 60$. I barely am OK spending that as it is - I'll wait for the 20$ sale. An all-you-can-eat netflix model might be viable, but has it's own issues especially in terms of dev kickback.

Oh and then there's the worst part about streaming. Like netflix, we can have content that is pulled. If a dev can't pay the server costs anymore, poof! What if I'm midway through a game, and the dev goes bankrupt? At least with a downloaded app I can still finish.

 

Phone infrastructure won't support it because it's going to be wifi only until 5G rollout is complete and even then, most people won't have 5G. Those that do will burn through their entire data plan in no time at all. This is no more a game changer than literally any other attempt at this.

 

CPU and GPU power usage is definitely not light, but neither is the radio signal that the phone uses to connect to its mobile network. Below is an older article (Summer 2010) detailing power consumption in a smartphone.

https://www.usenix.org/legacy/event/atc10/tech/full_papers/Carroll.pdf

Below is a snippet of some of the use case examples. Video playback aggregate power consumption was about 450 mW while the power required to drive the GSM mobile network radio signal for a normal phone call was more than double at about 1,050 mW. Wireless signals are massive power hogs and for this situation would need to be running non-stop.

Spoiler: Use Cases

 

The cost of the games also needs to be covered. The sub price can't just go to the hardware.

 

From a purely game dev or game marketing perspective, Google Stadia has quite a lot interesting takes: fully controlled environment + full debug data + anti-cheating + seamless youtube streamer recommendation.

From a consumer right, game dev independence, revenue sharing, game preservation perspective, Google Stadia is likely an absolute nightmare: instant global takedown and network dependence + limited partnership + existing subscription model aren't doing as well as you would like + no local copy.

Personally, I wouldn't take the former when I lose the latter.

(OK unless Google gives me a boat load of money to development a game for them... But I would still make more indie games on my own, really.)

 

I for one cannot wait for the future of indie games where I get paid $0.006 per play session. And the best part is I get to help youtube "creators" connect with their audience. Yay!