Server based game augmentations. The transition to cloud. Really possible?

[Nesh]

So, MMOs and online free-to-play games are a failed business model because there aren't enough people with stable Internet connections? Remember when Counter-Strike bombed because people were playing on dialup and broadband was new?

Again, the stability of people's Internet connections is besides the point. The thread is asking whether it's possible to augment a game with cloud computing. Not whether it's a feasible business model.

You are the one who went off topic because I did not talk about "market success"
MMOs and online free-to-play are successful because they are select titles that target a specific group of people that have access to good connections and love playing these specific games (and thank also internet cafes back in the firstCounter Strike days). Which is much different from making EVERYTHING and EVERYONE cloud reliant in some way and not just those who play online games.

You can probably make it work for a limited group of people in certain territories, with certain internet quality or at an experimental sampling level at this stage.
But the big question is, is it technologically viable to support all the users who will own a cloud based console for any game they unpredictably run and have a consistent quality experience? If not? How? If currently the technological obstacles cant be solved, When?

The thread title asks "The transition to cloud. Really possible?" And for that to happen the right infrastructure has to exist from multiple sides including the internet providers
edit: if MMO and Free-to-Play counted as "transition to the cloud" we wouldnt have been making this thread discussion because we would have assumed we are already there

 

[Shifty Geezer]

No takers so far. Maybe a better example would spur more discussion.

One post in five is on topic. Too much noise to moderate against.

The flag could de done online I guess, with the flag model with associated collision geometry around it (flag pole at the least). You'd have to send the processed model over the internet. Capes would be far more demanding and they'd need to be processed locally.

It's actually kinda hard to find computationally expensive tasks that aren't fairly time-critical and in response to user actions. Lighting has been mentioned. Destructive environments could be computed locally and then 'baked' online to new meshes for efficiency, I guess. And the old world background AI, of course.

 

[Tap In]

...The thread title asks "The transition to cloud. Really possible?" And for that to happen the right infrastructure has to exist from multiple sides including the internet providers
edit: if MMO and Free-to-Play counted as "transition to the cloud" we wouldnt have been making this thread discussion because we would have assumed we are already there

I can completely understand your point but if we base our future on factors that are currently 100% true, we are not innovating, we are not pushing the infrastructure to grow.

There is a train of thought counter to your suggestion that projects that it is exactly pushing the boundaries of a service like this on the infrastructure which will actually push legislators and internet providers to make positive change in that regard.

It is inevitable to have better connected world for ALL aspects of life.

 

[Scott_Arm]

One post in five is on topic. Too much noise to moderate against.

The flag could de done online I guess, with the flag model with associated collision geometry around it (flag pole at the least). You'd have to send the processed model over the internet. Capes would be far more demanding and they'd need to be processed locally.

It's actually kinda hard to find computationally expensive tasks that aren't fairly time-critical and in response to user actions. Lighting has been mentioned. Destructive environments could be computed locally and then 'baked' online to new meshes for efficiency, I guess. And the old world background AI, of course.

I suppose unless we get something to talk about from Turn10, or another first party, it'll be pretty tough to get anything other than a vague idea of what might be possible.

 

[Nesh]

I can completely understand your point but if we base our future on factors that are currently 100% true, we are not innovating, we are not pushing the infrastructure to grow.

There is a train of thought counter to your suggestion that projects that it is exactly pushing the boundaries of a service like this on the infrastructure which will actually push legislators and internet providers to make positive change in that regard.

It is inevitable to have better connected world for ALL aspects of life.

I agree with what you say. But let me add that for some things to improve and reach a point of viability they need a certain gradual evolution and the proper push from the right segment of the market.

And to make my stance more understandable lets go back in time to the 16bit era of gaming. We could have been discussing in the 90's "transition to online gaming.Really possible"?

Well we know now it is perfectly possible, but there were many reasons why back in the day this wasnt possible on console even though it was tried out (i.e the Genesis IIRC) and kind of worked for a limited niche market but with problems. And there were reasons why online console gaming wasnt as popular and effective until the PS3 and 360 were released. Some things need their time to mature and the right market demand (and product to push that demand)

Now I am not necessarilly saying that transition to cloud wont be possible until 10 or 30 years from now are passed. Thats why I raise questions/points that are mostly food for thought.

If the XBone IS the right product to push the server based game augmentation to the point that all infrastructures are also pushed to make it work flawlessly within this generation (and more importantly soon enough) is the million dollar question.

 

[golem]

I don't think the X1 (I think I prefer XBOne but I am childish like that ) is a good thin client candidate. It's too hot and heavy to deliver the '$50 bucks and a sub = awesum grafix' that remote gaming will one day deliver. I'd envision an Xbox version of the OnLive thin client as you can make those for half nothing. In terms of investment in the back end to support that investing later is always better as the cost curve on compute density, power consumption and capital cost only goes down over time.

As others have said think of the X1 as more of a bridge to that future than enabling it now. When that future does arrive im sure they'll release a pure thin client to satisfy customers looking for a low entry point but until then they need to sell a normal console box.

 

[astrograd]

I suppose unless we get something to talk about from Turn10, or another first party, it'll be pretty tough to get anything other than a vague idea of what might be possible.

I don't wanna post my several thousand character post from TXB here verbatim, so I'll just link to it for anyone interested in reading through it. I lay out things as I see them, starting from the opposite tract you did. I assume an effectively infinite bandwidth, establish what I feel are the possibilities and what it means to the platform and gamers, and offer a couple illustrative examples. The concept is to assume bandwidth isn't an issue and work backwards from there to narrow the scope if need be.

Here's my post: http://forum.teamxbox.com/showpost.php?p=13967394&postcount=55

 

[golem]

It's actually kinda hard to find computationally expensive tasks that aren't fairly time-critical and in response to user actions. Lighting has been mentioned. Destructive environments could be computed locally and then 'baked' online to new meshes for efficiency, I guess. And the old world background AI, of course.

One of the things I would hope to see now that the cloud is available for all X1 consoles are systems like mini-MMOs/expanded-multiplayer.. perpetual states that are easy for individual players to spawn and join-- much like Minecraft servers without the need to dedicate customer hardware to hosting.

An example of something I'd want to play would be like a cooperative Skyrim, a huge territory where players can roam around the world at will-- inviting any number of friends, completing quests and changing the political and economic states of the world-- with the data store and world simulation run in the cloud and local simulation run on each user's xbox. When there are no users logged in i would imagine the world state is suspended and saved until needed again, then any simulation elements that need to be computed due to passage of time would be executed upon restarting the world. That would be extremely cool for me and something enabled by having easy access to the cloud available to developers who want to take advantage of it.

 

[ERP]

Let's throw out something.

There is a broad spectrum between what Gaikai do and running everything locally.
For example I could theoretically run my entire game on the server minus the graphics, and instead of streaming video, I stream object state.
That seems no less feasible to me than the Gaikai model. Unless you think streaming video is lot less data that streaming game state.
Since all the physics/AI etc is in the cloud you can use all the local resources for making it look prettier.

I have perhaps 40K a frame at 30fps on a 10Mbit link, that's a lot of state and the only state I need is the stuff immediately around me. Most online games manage with less that 1/10th of that.

And there are is again a lot of spectrum between that and running locally.

Data caps obviously become a huge issue as they do with Gaikai.

It's certainly interesting, and to be honest I haven to applaud MS for putting the infrastructure in place. Whether it gets significant use in the next few years remains to be seen.

 

[Mize]

True scientific investigation has no motive other than to discover the truth, with no emotional attachment to what that truth is. No-one should care whether it can or can't work - we should only be discussing how it may or may not work.

Here's why I see it not working. How can anyone sell the x1, plus a fractional piece of server sufficient to augment the x1 for the price of the x1? Also, is that a guaranteed fractional server or do your games slow down or lose detail when more users come online?

 

[Lalaland]

....An example of something I'd want to play would be like a cooperative Skyrim, a huge territory where players can roam around the world at will-- inviting any number of friends, completing quests and changing the political and economic states of the world-- with the data store and world simulation run in the cloud and local simulation run on each user's device

This or a giant neural net/learning machine for AI are things that might leverage the cloud effectively. I'd like to see something like a 'social' RPG where each character has their own sophisticated like/dislike AI that is updated based on user interaction (perhaps even with a natural language engine?). Neural Nets & learning machines are very memory and CPU intensive while also being relatively latency insensitive (cloud response takes too long? Just add a 'thinking' animation loop). Of course learning machines are notorious for doing the weirdest things and are a mare to debug but I'd love for RPGs to move away from the simple binary like/dislike seen today

 

[ThatGamer]

Its not dead infinitely but at this stage and God knows for how long, its problematic because internet downtime is not something that happens once or twice at some point of our lives.
And perhaps this ISNT the time to rely on cloud computing especially on games but later when internet connections are improved.

I know people are pointing out latency problems and problems with internet dropping. In the worst case scenario devs could just pause/exit the game and attempt to reconnect or they might experience lag. Similar to what many MMOs already do. Is it user-friendly? No, but does it make sense to abandon the whole idea because there might be times when the internet drops?

People will not like it but most people will deal with it. They already face the same thing with online games today.

Also it is important to remember that Internet speeds are always increasing, in 8 years who knows how fast it will be. Today you can get faster internet on your phone than most people had at home 8 years ago.

Broadband speeds are increasing on average 25% YOY. http://techcrunch.com/2013/04/23/ak...th-korea-japan-and-hong-kong-still-far-ahead/

I would also wager gamers on average have faster internet than the average person.

Not only that but the FCC is pushing for Gigabit internet in all 50 states by 2015. http://news.cnet.com/8301-13578_3-57564815-38/fcc-pushes-for-gigabit-broadband-in-all-50-states-by-2015/

The answer to what can be done in the cloud with regards to games will be different today then it will be in 5 years. Technology is always moving forward and the internet is always improving. So while devs might be constrained to a 10 mbit bandwidth, in 5 years who knows what they will have to work with.
http://news.cnet.com/8301-13578_3-5...r-gigabit-broadband-in-all-50-states-by-2015/

 

[Nesh]

I know people are pointing out latency problems and problems with internet dropping. In the worst case scenario devs could just pause/exit the game and attempt to reconnect or they might experience lag. Similar to what many MMOs already do. Is it user-friendly? No, but does it make sense to abandon the whole idea because there might be times when the internet drops?

People will not like it but most people will deal with it. They already face the same thing with online games today.

We have learned to face it in online games because online gaming was having its fair share of problems since its conception and improved from there. Hence we have grown to accept online for what it offers as it improves in time.

Unlike online gaming (which is just a part of the whole gaming experience), offline gaming was a steady and consistent experience since the inception of videogames. You put in the game and you play. Nothing can stop you or affect how the game performed. A perception that has been constantly enforced for half a century of consistent experience.

The reliance on cloud gaming for every game suggests taking the consistent portion of gaming (lets say a single player campaign), introduce in it the inconveniences found in online gaming and expect everyone to accept it like its ok because online was always having some issues. For me this doesnt make sense

And for the record I want to play what I like and enjoy it as much as possible, not force me issues that werent there in the past and tell me to deal with it
Dealing with problems does not equal fun and I want to have a consistent and enjoyable experience with my product.
Consistency is key in every good product and service
Admitting "people will not like it" equals admitting that there is a problem. It is even worse when you tell them to "deal with it" as if that makes it a non-problem

Also it is important to remember that Internet speeds are always increasing, in 8 years who knows how fast it will be. Today you can get faster internet on your phone than most people had at home 8 years ago.

Broadband speeds are increasing on average 25% YOY. http://techcrunch.com/2013/04/23/ak...th-korea-japan-and-hong-kong-still-far-ahead/

I would also wager gamers on average have faster internet than the average person.

Not only that but the FCC is pushing for Gigabit internet in all 50 states by 2015. http://news.cnet.com/8301-13578_3-57564815-38/fcc-pushes-for-gigabit-broadband-in-all-50-states-by-2015/

The answer to what can be done in the cloud with regards to games will be different today then it will be in 5 years. Technology is always moving forward and the internet is always improving. So while devs might be constrained to a 10 mbit bandwidth, in 5 years who knows what they will have to work with.
http://news.cnet.com/8301-13578_3-5...r-gigabit-broadband-in-all-50-states-by-2015/

Be aware that I did not rule it out completely for the future so that portion of your post is irrelevant to what I have stated earlier

 

[ThatGamer]

We have learned to face it in online games because online gaming was having its fair share of problems since its conception and improved from there. Hence we have grown to accept online for what it offers as it improves in time.

Unlike online gaming (which is just a part of the whole gaming experience), offline gaming was a steady and consistent experience since the inception of videogames. You put in the game and you play. Nothing can stop you or affect how the game performed. A perception that has been constantly enforced for half a century of consistent experience.

The reliance on cloud gaming for every game suggests taking the consistent portion of gaming (lets say a single player campaign), introduce in it the inconveniences found in online gaming and expect everyone to accept it like its ok because online was always having some issues. For me this doesnt make sense

It seems you are making many assumptions. I don't think anyone is implying that every game will be forced to use the cloud. I certainly wasn't implying that. Also games like Diablo 3 that can be played single player have shown that you can be a big success even if you are forced to be online in single player games. I am not saying it is the best/worst choice. That would be up to the devs/publishers and their unique circumstances.

I am simply saying that there is evidence consumers will accept it. You may not personally like that but that is irrelevant. Businesses will not make decisions based on what you like. Also remember this is just illustrating a worst case scenario. They could degrade gracefully as well.

And for the record I want to play what I like and enjoy it as much as possible, not force me issues that werent there in the past and tell me to deal with it

I said in the worst case scenario. Maybe there won't be any issues at all. Where did I tell you to deal with it? I am not trying to convince you to accept or deal with anything. I am showing that devs/consumers have faced similar problems in the past. Read what I actually said:

"People will not like it but most people will deal with it". There is evidence of this already. Many people deal with lag/disconnecting in online games already. I didn't say that all people would or that you specifically would need to deal with it... I am simply stating that people are already faced with these problems all around the world.

Dealing with problems does not equal fun and I want to have a consistent and enjoyable experience with my product.
Consistency is key in every good product and service
Admitting "people will not like it" equals admitting that there is a problem. It is even worse when you tell them to "deal with it" as if that makes it a non-problem

I didn't tell you to deal with it. I was illustrating the worst case scenario. Where did I say it was a non-problem? You are making a lot of assumptions...

I am simply stating that even if it isn't perfect many customers will accept it because we have evidence that similar situations have been accepted in the past.

Be aware that I did not rule it out completely for the future so that portion of your post is irrelevant to what I have stated earlier

Yes, but these consoles will be around for years. Also to base our conlusions on what is available today is illogical because the possibilities will improve as the consoles age.

 

[MrSpiggott]

I think we can all agree that large open world organisation is perfect for cloud processing, but the argument over whether its going to make a visual difference when compared with PS4 seems to be going around in circles, and will probably continue to be argued about for many years into the machines life.

Lets take physics as an example and try to put some numbers behind the questions.

First I'm going to make some very broad assumptions.
1. There will be more than enough servers to cope with any demand. If everybody jumps on with Halo 5 on launch week it'll handle it.
2. Sony never offers cloud processing
3. The minimum internet bandwidth requirement for this feature is 7Mb/s. I've suggested this value as it's the minimum spec' for high quality HD with netflix. I would have imagined they've carried out a lot of research and come up with this figure as a sweet spot to maximise profit. (The various figures I've read for countries average bandwidth don't come with any methodology, or the type of averages used, so I think netflix's number is a safe guess.)

There are going to be five stages to cloud processing.
A. The time taken for Xbox to process information ready to be sent to the cloud.
B. Connection latency + time taken to send a minimum amount of useful data to be worked on by the cloud @ 7Mb/s.
C. Cloud rendering time.
D. Connection latency + Returning data @ 7Mb/s.
E. The time taken for the xbox to incorporate this data and render the frame.

Can anybody here add any numbers for the above.
What are the CPUs being used in the Microsoft farms? I'm getting the impression the servers are CPU only with no GPGPU support, but haven't been able to confirm it. Is that correct?
What are typical latencies to server farms within 50 miles? (I've unintentionally collected about a years worth of ping data from between my PC, ipad and phone, and my internet provider, but have no way of vouching for its reliability.)
What is the hypothetical minimum amount of data that would have to be sent to calculate a building collapsing?

So to be carried out in one frame would mean A+B+C+D+E=33ms, but I think everyone's agreed on that this can't happen with any reliability, but what about a 10fps effect update?
Is 66ms+C (where C=33ms-A-B-D-E) going to be long enough for a CPU running various VM to calculate the collisions of yours and other Xboxes buildings collapsing?

If it is how does that compare to PS4? It's got an extra 6 CUs so can afford to use them exclusively to calculating that building collapse, and its going to have a longer time to do it in. 66ms+B+C+D.
Now the question is can the physics calculations be carried out on these 6CUs? And if it can will it be as quick as the same work being carried out on a CPU in a server?

If the amount of work carried out by cloud processing can equal or better that of 6CUs then it will one day offer noticeble improvements in visuals, for physics at least.
Any flaws in my logic?

 

[ninzel]

First thing I thought when I heard them talking cloud for gaming was "accessory " and isn't the rule of thumb that developers code only for what they can count on?
That's just my simple assessment of its potential.

 

[astrograd]

Devs can count on having 24 CPU cores and 32GB of memory in the cloud. That's what MS said.

 

[ninzel]

Devs can count on having 24 CPU cores and 32GB of memory in the cloud. That's what MS said.

I'm thinking of from the opposite side. The player. Doesn't the power of cloud depend on both server and client? How can devs count on the clients connection?
Edit: So your playing your shiny new game and for some reason your internet connection drops or slows down and all of a sudden features start dropping off? Or because your internet speed fluctuates the game pauses and a message comes up saying" please wait while we buffer the cloud"?

 

[Betanumerical]

Devs can count on having 24 CPU cores and 32GB of memory in the cloud. That's what MS said.

I haven't seen any place point out the memory part, can you give us a link. It wouldn't make sense in any regard really considering its 3x the XBONE specs and the XBONE has 8 cores and 8GB of memory it'd be 24 cores 24 GB of memory.

 

[astrograd]

I haven't seen any place point out the memory part, can you give us a link. It wouldn't make sense in any regard really considering its 3x the XBONE specs and the XBONE has 8 cores and 8GB of memory it'd be 24 cores 24 GB of memory.

Bah, yeah nvm. Not sure what I was thinking. It's be 24 of each presumably. Maybe even 18 cores and 15 GB, depending on what they mean by x3 X1 resources. I was thinking 32 cores overall and typed in 32 for some reason, heh.