Game Streaming Platforms and Technology (xCloud, PSNow, GeforceNow, Luna) (Rip: Stadia)

[Malo]

By the same token, doesn't the concept of client and host on the same massive network bring opportunities a regular server-client can't cope with? The idea of a mass fantasy melee battle with 1000 players is alluring but only really possible on a high capacity LAN. If a game is designed around that latency to the client for the input, there's opportunity for these services to provide something not previously possible.

 

[iroboto]

By the same token, doesn't the concept of client and host on the same massive network bring opportunities a regular server-client can't cope with? The idea of a mass fantasy melee battle with 1000 players is alluring but only really possible on a high capacity LAN. If a game is designed around that latency to the client for the input, there's opportunity for these services to provide something not previously possible.

Probably, that's a programming problem though and I'm not sure how many studios would be willing to invest in that, I suppose some. But a lot of budget is going to be spent there and not on graphics, etc.
Example 1: See crackdown 3

 

[Shifty Geezer]

By the same token, doesn't the concept of client and host on the same massive network bring opportunities a regular server-client can't cope with?

This possibilties are probably more illusionary. That many players in a battle, you won't see them all. You'll only ever see maybe 32 v 32. MAG had 256 player matches and it was no different to playing a 32 player match. With no communication or connection with teams on other parts of the maps, their efforts and objective were effectively meaningless, and you could just script virtual battles outside of a 32 v 32 match as if other players were involved. You're also limited in what creative elements your players can have. Starhawk had players able to build bases, but with zero coordination it was just a mess. Even if powerful servers would allow things like 1000 mages to cast spells to terraform and summon creatures and control the weather, it basically become a completely chaotic, unplayable mess. Such a 'mage wars' game would maybe have to limit to 5 v 5 as the only sane way to balance everything that's happening to actually have a fun game.

 

[AlNom]

Google can amortize the cost of the hardware on multiple subscribers because they won't all be playing at the same time.

Further, with virtual GPUs a single GPU can run mulitple games simultaneously. Google can run 5 instances of Castle Crasher on the same GPU that runs one instance of Medal of Duty 8.

Cheers

Indeed. I wonder how that would work in practice for various MMOs that are not particularly sensitive to latency as well, and are perhaps not so demanding as to be able to run multiple instances on a particular setup?

 

[Malo]

This possibilties are probably more illusionary. That many players in a battle, you won't see them all. You'll only ever see maybe 32 v 32. MAG had 256 player matches and it was no different to playing a 32 player match. With no communication or connection with teams on other parts of the maps, their efforts and objective were effectively meaningless, and you could just script virtual battles outside of a 32 v 32 match as if other players were involved. You're also limited in what creative elements your players can have. Starhawk had players able to build bases, but with zero coordination it was just a mess. Even if powerful servers would allow things like 1000 mages to cast spells to terraform and summon creatures and control the weather, it basically become a completely chaotic, unplayable mess. Such a 'mage wars' game would maybe have to limit to 5 v 5 as the only sane way to balance everything that's happening to actually have a fun game.

That's limiting yourself to concepts of existing games. I'm thinking of Total War-scale fantasy combat where armies might face off against each other on a field of battle where there's literally hundreds of combatants on each side running at each other down hills into a mass melee Braveheart style. Each combatant is an AI bot that could be taken over by a player with potentially everyone being a real person. Would it be a viable game in reality? Probably not, but for reasons I would rather not be due to technical reasons where cloud concepts would solve.

Just trying to think outside of what's accepted limitations.

 

[JoeJ]

I'm thinking of Total War-scale fantasy combat where armies might face off against each other on a field of battle where there's literally hundreds of combatants on each side running at each other down hills into a mass melee Braveheart style. Each combatant is an AI bot that could be taken over by a player with potentially everyone being a real person.

That's pretty much what i would expect from an innovative cloud gaming service. The idea is very present since the days of Command & Conquer and first ego shooters, but nobody ever touched it yet?
From a game design perspective the problem seems 'how can i motivate the FPS players to follow the orders of RTS players'? Would they obey? Or just cause chaos for fun?
From a technical perspective, the promise 'each player gets its own machine' seems to hint they do not have something like this in mind, showing particle mGPU simulations instead?
Although - joining CPUs would be more interesting here and maybe they do not give the core count for this reasons? Can they form a large server that does all physics and game logic, and each player only gets a small virtual client CPU for processing local input, camera, audio etc.?

 

[Metal_Spirit]

Joej... Is that a good thing?



If you can get on a 30 euros Chromecast device the same result as on a 1500 euros PC, what would you spend money on such a PC or any other hardware?

This will lead to a decrease in sales in gaming hardware, a decrease in profits, in search and development, and a cost increase in overall hardware.

Now... This will afect all companies that sell hardware. Prices will go up, software and hardware companies will sink. This will mess with the world economy!

Countries will suffer, wars will be declared.

OMG... Google just announced the end of the world

 

[pcchen]

Although - joining CPUs would be more interesting here and maybe they do not give the core count for this reasons? Can they form a large server that does all physics and game logic, and each player only gets a small virtual client CPU for processing local input, camera, audio etc.?

If you use a huge closely connected supercomputer to run games, it's possible to share even physics simulations across players so it'd be cheaper, and it could open a lot of new possibilities (for example, a real destructible terrain in a huge battlefield). However, I doubt that Google will be doing something like this anytime soon, as interconnection is generally the most expensive part of a supercomputer.

 

[mpg1]

Joej... Is that a good thing?



If you can get on a 30 euros Chromecast device the same result as on a 1500 euros PC, what would you spend money on such a PC or any other hardware?

This will lead to a decrease in sales in gaming hardware, a decrease in profits, in search and development, and a cost increase in overall hardware.

Now... This will afect all companies that sell hardware. Prices will go up, software and hardware companies will sink. This will mess with the world economy!

Countries will suffer, wars will be declared.

OMG... Google just announced the end of the world

The data centers still have to use hardware...

 

[JoeJ]

OMG... Google just announced the end of the world

Haha, yes, or better: World ownership, because all the remaining hardware is theirs then, and we will live an almost virtual life within their kind hands.

If you use a huge closely connected supercomputer to run games, it's possible to share even physics simulations across players so it'd be cheaper, and it could open a lot of new possibilities (for example, a real destructible terrain in a huge battlefield). However, I doubt that Google will be doing something like this anytime soon, as interconnection is generally the most expensive part of a supercomputer.

They should do it a quickly as possible. I guess it becomes hard to make profit otherwise, if they want more than moving twitch users to YT.

 

[Shifty Geezer]

Each combatant is an AI bot that could be taken over by a player with potentially everyone being a real person.

But what does that bring? It's one on one with a load of stuff happening around, or a few on few. What's the difference between that and 16 v 16 with a bunch of cosmetic AI fighting around, or even where you could leap into any AI bot as you suggest?

From a game design perspective the problem seems 'how can i motivate the FPS players to follow the orders of RTS players'? Would they obey? Or just cause chaos for fun?

Precisely. The bigger the player count, the more trouble you have coordinating everyone. 1000 people in a game will mean a number of jerks and arses, unavoidable with those kind of numbers, and a lot of solo players. We could have had massive games before on servers, but I think it hasn't happened because devs realised it just wouldn't work in all their cases. The CODs of this world stuck to 32 v 32 for good reason despite 128 v 128 being proven.

Looking at the latest BR craze, how much different would it be having a map 10x the size with 1000 players instead of 100? For the people playing, it won't feel any different - you'll have the same density of encounters. If you wanted to make it feel denser, you could just shrink the map instead of needing to add more players. The only time having huge populations would feel different is when the number engaged at any moment is massive, at which point you tend towards chaos. Picture Apex or COD with 10x as many people shooting at you at any moment - it'd just be too much. But then if the numbers are spread out enough to avoid chaos, than the connection between players is reduced and the impact of having huge populations pretty much mitigated, such that you could simulate a larger universe based on the actions of a smaller team and it'd feel exactly the same to them as if there were hundreds of other human beings being affected. In MAG, your bunch of 32 players would have an objective, and another 32 players on you team would have another, and you'd see on the map that they completed their thing. Doesn't need 32 humans to achieve that - you could just ping the map based on how well your squad was doing.

The only benefit I can really think of comes with streaming, with 1000 people playing the same match potentially resulting in a very strong following in the case of epic matches. But I'm really struggling to see any situation where a large population results in a meaningfully different game that isn't chaos. What's likely more useful is massive persistent storage, so thousands of people can build/shape a world and see each others' impact. I suppose a 1000 player Tomorrow Children type thing might be interesting, except you then have the problem with each player wondering about their value to the game. Without them helping...nothing's going to really be worse off, so why not just screw around moving everyone's stores where they aren't needed (which has a sense of impact and agency, and perhaps that's why people are jerks? I might have just stumbled onto a key game-design philosophical concern!).

 

[Malo]

But what does that bring? It's one on one with a load of stuff happening around, or a few on few. What's the difference between that and 16 v 16 with a bunch of cosmetic AI fighting around, or even where you could leap into any AI bot as you suggest?

As I said, it's not a game concept that would be worth it but there are thousands of people more capable of coming up with exciting concepts than me. My point was that the cloud concept can bring ideas that are no longer restricted by the client-server limitations.

 

[AzBat]

But I'm really struggling to see any situation where a large population results in a meaningfully different game that isn't chaos.

1 vs 1000

Tommy McClain

 

[JoeJ]

The only benefit I can really think of comes with streaming, with 1000 people playing the same match potentially resulting in a very strong following in the case of epic matches. But I'm really struggling to see any situation where a large population results in a meaningfully different game that isn't chaos.

I think the benefit is quite large, considering the streaming audience can spot the ingame Ads with more attention while watching
But more than this, this attention can also be used to limit the chaos. Players knowing being watched might be after reputation. Reputation could even be the currency for upgrades. So causing chaos -> worse guns, less badges on your skin.
In a RTS > FPS scenario, the less reputated players will also get a worse start position ending up as cannon fodder, while the good guys get the cool missions like stealing the flag.
I think it can work if the chaos guys are already taken into account in game design, and they can be jerks on purpose - still having fun and value for the game.

Actually BR is already an example for this. When i played Quake3 Clan game mode (team last man standing), i wondered: Would this be fun with no players? No - too much waiting time after death.
But now with BR the same thing is very successful. It even is a 'new genre', and new genres is what we need the most, and what we should expect form new technology. New genres could even help with further growth to prevent my pessimistic prediction of a downfall in gaming industry.
I think streaming is one indication of an upcoming downfall. It did not work so well for the music business. Just moving existing games to the new platform will only weaken the industry - we need to give those platforms new kinds of games, not just an easy way to play them.

I suppose a 1000 player Tomorrow Children type thing might be interesting, except you then have the problem with each player wondering about their value to the game.

I don't know TTC, but if we take Minecraft as an example, or one step further - a similar game where the players are also the content creators, then we have the final single remaining game we likely are all heading to.
Think of a game where players create their own minigames like in Dreams but in a consistent large world like Minecraft with consistent physics and ecology. They will also form larger communities to create larger games / experiences / whatever. There are always enough good level designers / artists / script coders within a large community.
Again reputation could be used as a currency to keep the whole thing working, together with laws of physics, finite transformable resources as in Minecraft, or whatever else. I fdone right, the game would work without any scripted missions, without telling the players what to do, without never ending afford to add new content to the game because the users do it better than we could do.

Now i don't know if i want all this to happen, but it will. At first there is a lot of work for us to do to get there, but at some point we will be no longer needed?
You may think: 'No - users can't do this, they'll just create mediocre experiences and chaos', but i doubt it because the best games are (or were) those with user generated content.

 

[Jay]

Upto 10 person clan with 100 teams all at once?
Tetris 1 vs 999 matches .

I finally got to watch the Stadia Google keynote.
It will get new current gen games ported, and next gen also. Possibly some older but still popular games.
It has the power, to scale to 2units when next gen comes out for games that require more than the 1U.

Xcloud has BC games, and will only have a limited set of current gen games available to play.
My reasoning that what should be its biggest selling point of XO catalog of games, will be missing is simply down to performance.
Games running at 25-30, and 40-60 fluctuating all over the place will impact on perceived latency and give a really bad experience.
So I think they just won't be put on there.

MS will say its just a beta, while sadia is full release.

Edit : I also expect Google to money hat some exclusives, timed and content. Does the phrase money hat have negative connotation? If so not meant that way

 

[Shifty Geezer]

Does the phrase money hat have negative connotation? If so not meant that way

Yeah, money hatting comes with it a sense of snatching up games that'd otherwise come to other machines. That's a little different to investing in 2nd party titles and new IPs. The alternative phrase would be to say, "I also expect Google to invest in some exclusives, timed and content," which is a more constructive act rather than pulling games and content away from other platforms. Timed exclusives are always money-hatting and negative connotations.

 

[Jay]

Yeah, money hatting comes with it a sense of snatching up games that'd otherwise come to other machines. That's a little different to investing in 2nd party titles and new IPs. The alternative phrase would be to say, "I also expect Google to invest in some exclusives, timed and content," which is a more constructive act rather than pulling games and content away from other platforms. Timed exclusives are always money-hatting and negative connotations.

Yea, I would prefer for them to go about it in the nicest possible way.

But, if they Epic some stuff I wouldn't loose any sleep. Rather wasn't permanent exclusives though.
They need to get (exclusive) content as quickly as possible to give people a reason to 1 give a dam, 2 to come back to it.
The general multiplats will find their way onto there.

I doubt streaming itself will be enough to get gamers to give up console, or invest in console and a different streaming provider.

 

[iroboto]

Yea, I would prefer for them to go about it in the nicest possible way.

But, if they Epic some stuff I wouldn't loose any sleep. Rather wasn't permanent exclusives though.
They need to get (exclusive) content as quickly as possible to give people a reason to 1 give a dam, 2 to come back to it.
The general multiplats will find their way onto there.

I doubt streaming itself will be enough to get gamers to give up console, or invest in console and a different streaming provider.

They may not have to. it's not a zero sum game and Microsoft is more than willing to concede exclusivity in favour of opening up and gaining profits in another way.
IE: XCloud or Stadia could be infrastructure sold to both Sony and Nintendo should they choose to. It's definitely not impossible if your servers support
a) a higher feature set
b) a more power
There's no doubt in my mind that if MS could sell XCloud to either competitor it would cause them to leave Amazon and come onto Azure. That's already a big win because that type of service is guaranteed massive profits and little risk of not getting paid. Games are still relatively speaking risky and spikey business.

The challenges really comes down to tools for developers. Which I would properly assume that MS is ahead of Google in this manner as this is their core business.

 

[Jay]

They may not have to. it's not a zero sum game and Microsoft is more than willing to concede exclusivity in favour of opening up and gaining profits in another way.
IE: XCloud or Stadia could be infrastructure sold to both Sony and Nintendo should they choose to. It's definitely not impossible if your servers support
a) a higher feature set
b) a more power
There's no doubt in my mind that if MS could sell XCloud to either competitor it would cause them to leave Amazon and come onto Azure. That's already a big win because that type of service is guaranteed massive profits and little risk of not getting paid. Games are still relatively speaking risky and spikey business.

The challenges really comes down to tools for developers. Which I would properly assume that MS is ahead of Google in this manner as this is their core business.

You need to think like a rich person or business, there's never enough money.

So even if they sold the platform to be rebadged by other companies it doesn't mean they wouldn't want their own to be as successful as possible.
Selling it is just another revenue stream.

Could get to a point where it just makes more sense to only sell the cloud service rather than to provide service to gamers. But I don't think any of them are thinking like that in the near future.

I just don't see any of them not thinking content especially exclusive isn't important. Ms tried that for a few years, and have now pivoted.

 

[iroboto]

You need to think like a rich person or business, there's never enough money.

So even if they sold the platform to be rebadged by other companies it doesn't mean they wouldn't want their own to be as successful as possible.
Selling it is just another revenue stream.

Could get to a point where it just makes more sense to only sell the cloud service rather than to provide service to gamers. But I don't think any of them are thinking like that in the near future.

I just don't see any of them not thinking content especially exclusive isn't important. Ms tried that for a few years, and have now pivoted.

I'm not referring to content exclusivity but service exclusivity.
MS knows that they will never be about to own 100% of the market. IE: RDR2 sales are divided among PS4 and XBO for instance. They will never see any profit of the ones sold on Playstation. If you make streaming exclusive, perhaps you'll maybe get a couple more percentage points on the sales because of switch over as being a driver, but you'll see significantly more profit if you support the Playstation streaming, and now you're taking a small percentage of all sales in the whole market. They just went from 100% of 20/30% to 100% of 20/30% and X% of the remaining 70%.

MS will never profit from Sony exclusives or nintendo ones, this avenue of supporting streaming is ideal here for completely new revenue and additional saturation of their assets.