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

[Scott_Arm]

This is true. But I cannot help but think that a bunch of things that are being suggested could be done easier and quicker on the local machine, a lot quicker if they were done well. It seems like the cloud would be useful for things that the player does not directly interact with, or things that require a large of amount of data (market simulations?, player interactions) but I swear the same things are being suggested over and over again and nearly all of them see to involve player interaction I cannot see this working due to latency issues. Something that no one has addressed yet and something I do not think will improve anywhere near the rate that bandwidth will.

There is definitely a need to pick and choose what goes to the cloud very carefully, if the cloud turns out to be a realistic place for game computations at all.

And you are correct about latency. In fact, latency has been getting worse as bandwidth has increased. I read about Buffer Bloat a while back. Hopefully some academics and the telecom industry will come up with some viable solutions that get adopted reasonably quickly.

 

[Shifty Geezer]

I've just realised another potentially serious factor affecting cloud viability, and that's when downloading content. If you have background downloading active, whether downloading the game you are currently playing or some other game or content, your hardware is going to be saturating the BW. Unless that is throttled, QoS cannot be guaranteed to any level, seems to me. You can't even be sure of 1 Mbps, 30 kilobytes per frame, of content to be available at a given moment. Things like just-in-time calculations as astrograd suggests become even more implausible, unless all of a few kb. Then the cloud service can interrupt any local downloads for the window. It'll of course have no control over bandwidth used elsewhere by the local network. Things like video chat will also impact available BW.

 

[Shifty Geezer]

There is definitely a need to pick and choose what goes to the cloud very carefully, if the cloud turns out to be a realistic place for game computations at all.

And you are correct about latency. In fact, latency has been getting worse as bandwidth has increased. I read about Buffer Bloat a while back. Hopefully some academics and the telecom industry will come up with some viable solutions that get adopted reasonably quickly.

I thought latency was improving but not as fast as BW, so relative latency was increasing? Faster routing tech and servers should decrease time taken to handle network traffic, although of course more traffic has the opposite effect, and if amount of traffic is increasing faster than technology abilities to deal with it, you may be right.

 

[patsu]

This is exactly what I am curious myself: either MS, the publisher/dev or the user/gamer has to pay for the cloud resources.

I think many developers are asking the said same question.

Eh ? By this time, basic questions like this should already been addressed. Unless the cloud is only used by the first parties today.

come on now... this is not a mature system yet. Be reasonable... all the servers are not even built yet... they are probably still working on the API's and yes they said First party will lead the way and they will give devs the tools and hope they use their imaginations to grow it


this is a system that will grow and mature over time... it is new in its infancy... Al of the people that are pointing this out now are misunderstanding that this is a 2-5 year cycle of exploration and MS is taking the leap for us to see where it goes.

It's called progress and pushing the envelope of technology. It might fail but hell I have rarely seen so many technically minded people be so resistant to an idea of pushing boundaries, exploring what was previously thought to be impossible

Huh ?

http://beyond3d.com/showthread.php?p=1740431#post1740431

If they are at this early stage, it would be no different from DUST514 using their own servers *yesterday*

The cloud services are "embedded" into the game.

They will need to avoid this type of mixed/conflicting messages.
Did they confirm dedicated server gaming for XBL subscribers ?

Launch is only 6 months away you know ? And if they haven't worked out the business case, and the servers are not built yet, then it doesn't look good. It sounds more like the first parties' private effort.

they have lots of servers deployed in general, 300k assigned for LIVE they said will be by launch

as suggested by a couple of the articles yesterday the move engines are possibly there to help allocate cloud/local data.

not sure where you are getting that they are not ready from what I'm saying? its just premature for you to judge that its not going to work yet. All I'm saying is it's 6 mos away til launch and then I'd guess another 2 years til you would expect to see it developed to the point you (and Ms actually) are projecting

Read the above discussion. If MS still haven't told their partners whether and how much they are going to pay for the cloud servers by now, then it is very late. Development has already started months ago. Even if the infrastructure is not mature yet, the budgeting and business planning has to happened already. What's so unreasonable about asking for how much devs are going to pay (if any) for the infrastructure ?

As for maturity of the cloud platform, earlier on, you mentioned that they have been deploying servers to prep for XB1 cloud, and now you said the servers are not built yet. Where did you get your information from ? For such a large scale operation, it's usually planned and executed over years. It needs to be integrated with existing XBL infrastructure and tuned. It's not just the boxes.

And I didn't say it's not going to work. Check my post history. I even suggested multiple examples where cloud can help. >_<

I merely pointed out that the MS execs messed up their own launch. They had random execs spouting unverified info to different press. e.g., The earlier info about 3 virtual XB1s were later clarified to be "just" the CPUs.

they said that each game would get it's own server(s) yes

That's better. These basic dedicated server benefits are more tangible than the other advanced stuff, at least in the early stage.

 

[patsu]

What if developers started thier own cloud servers and ditched microsoft,sony and nintendo?
Think about it.

EA network gaming service.
Activision network gaming service
Sega-Sammy network gaming service.
Capcom network gaming service.

Access to their whole back library and currently available games
.
Developers don't even have to maintain the cloud network they can just contract out a private cloud provider and just focus on the creating the content.
They can then warehouse thier content or lease them out to private cloud gaming publishers.

Yes, they could pay Microsoft, Amazon or Google to do that. Not sure how the financials work out. Probably better than trying to build your own data centers across the world, especially when they have no internal expertise in doing that kind of thing.

Yes, that's the status quo. MMOs and network games today/yesterday already roll their own server farm on assorted facilities. Depending on the nature of the game, they may be doing assorted simulation, AI on the servers.

Microsoft's move here should be *above* the cloud infrastructures. It should be seen as an aggregation of many parts:
* XBL paying user community (and their points)
* Exclusive software library and media contents
* Secure/DRM'ed server infrastructure streamlined for entertainment services
* Accessibility via Kinect
* Spare server capacity for assorted future work

While the tech fora are debating the last point, it is the aggregation that is more interesting to the business folks. The used game sale (or no sale) mechanism may be part of a larger (re)pricing plan.

On technical subjects, the cloud servers should be applicable to any "large enough grained" async tasks. e.g., A game character typically walk/run slowly, they can also preload dynamic world content (user and developer generated) from the next area from the cloud. Both consoles have large memory. If the world is pretty static, then just preload them on the BR disc, or download them periodically from the servers. That's kinda what Playstation Home does albeit in a clumsy way.

I would be interested to have DF talk to WoW, FFXIV, Planetside 2 and EVE + DUST514 to give an overview of their technologies from rendering to world setup; plus future plan. It would make the discussion more tangible.

EDIT: The WWII plane sim MMO, War Thunder (?), may be interesting too.

 

[Scott_Arm]

I thought latency was improving but not as fast as BW, so relative latency was increasing? Faster routing tech and servers should decrease time taken to handle network traffic, although of course more traffic has the opposite effect, and if amount of traffic is increasing faster than technology abilities to deal with it, you may be right.

My understanding was that overall latency is getting worse. It has been a while since I reviewed the material

http://en.wikipedia.org/wiki/Bufferbloat
http://www.bufferbloat.net/projects/bloat/wiki/Wiki

 

[Billy Idol]

I've just realised another potentially serious factor affecting cloud viability, and that's when downloading content. If you have background downloading active, whether downloading the game you are currently playing or some other game or content, your hardware is going to be saturating the BW. Unless that is throttled, QoS cannot be guaranteed to any level, seems to me. You can't even be sure of 1 Mbps, 30 kilobytes per frame, of content to be available at a given moment. Things like just-in-time calculations as astrograd suggests become even more implausible, unless all of a few kb. Then the cloud service can interrupt any local downloads for the window. It'll of course have no control over bandwidth used elsewhere by the local network. Things like video chat will also impact available BW.

Exactly what I wondered a couple of posts ago. And don't forget the potential influence of the promoted simultaneous apps...can you simultaneously Skype on your small app window? How do such apps eat BW and influence potential cloud usage?

 

[Arwin]

Currently both consoles also throttle or if necessary stop background downloads when you're playing a multiplayer game online, so this is not necessarily rocket science ...

 

[Scott_Arm]

Doesn't seem impossible that they've have QoS built into the box so game packets are always handled first. Good question though.

 

[Rangers]

My cable connection just got upgraded from 15 Mbps to 30 Mbps for $5 less per month. That's right, they subtracted $5 to double my speeds

I toyed with the 107 Mbps they offer, but that's like $85 a month.

Clouds a-comin!

 

[astrograd]

I've just realised another potentially serious factor affecting cloud viability, and that's when downloading content. If you have background downloading active, whether downloading the game you are currently playing or some other game or content, your hardware is going to be saturating the BW. Unless that is throttled, QoS cannot be guaranteed to any level, seems to me. You can't even be sure of 1 Mbps, 30 kilobytes per frame, of content to be available at a given moment. Things like just-in-time calculations as astrograd suggests become even more implausible, unless all of a few kb. Then the cloud service can interrupt any local downloads for the window. It'll of course have no control over bandwidth used elsewhere by the local network. Things like video chat will also impact available BW.

True, so turn off those OS features. Problem solved. The game should be plenty capable of discerning how much download bandwidth it alone can use in order to scale the cloud applications.

I think we should look at Mary Jo's article on MS's Orleans stuff. Seems like it could very well be MS's testbed for their cloud concepts. Her article on it has lots of info that may be useful.

 

[warb]

20+Mb/s is he minimum/cheapest available from cable here. In 5-10 years time we'll probably all be on 100Mb/s and up.


I imagine games will stop or throttle background downloads when the best connection is required for cloud stuff, as some apps and games do now. Anything that can be done on the cloud would need a simpler place holder on the local system.

 

[-tkf-]

True, so turn off those OS features. Problem solved. The game should be plenty capable of discerning how much download bandwidth it alone can use in order to scale the cloud applications.

I think we should look at Mary Jo's article on MS's Orleans stuff. Seems like it could very well be MS's testbed for their cloud concepts. Her article on it has lots of info that may be useful.

And we ask the rest of the house to turn of whatever they are doing because the cloud is busy making calculations?

"Can i watch netflix? NO! i need to play this awesome game that runs in the cloud"
"STOP WATCHING YOUTUBE I NEED TO PLAY"
"Close your browsers, i need to play"

 

[astrograd]

And it's slow CPU power compared to the onboard CPU for all the reasons mentioned a million times.

I haven't seen that mentioned. You know the specs of thse CPU's? Link?

Cloud computing is very good at handling extremely many users at the same time, but they do it slow compared to local resources.

It's generally good for anything involving loads of moving parts. That could be lots of players in an MMO or it could be a robust physics and dynamics simulation of a building collapse in an industrial application.

As i read this thread i am pretty confident that the number of games that will use this is limited to 1st party games from either Nintendo,Sony or Microsoft.

I can see exclusives in general being a target. I'd agree that for a while it may not be suited for multiplatform games. Then again, it could also very well end up being something that works well enough to bring exclusives to the X1 too for all we know. It's gonna be up to MS to really demonstrate these capabilities. Could be why their first party investment is so substantial this time around. If they can even get cloud stuff working in legitimately impressive ways in each major genre that'd be a strong start.

I think we still lack a well written example of cloud computing lifting work from a console game that makes it worth the effort.

It was announced only a week ago and we don't actually have any direct info on the games for X1 anyhow beyond vague announcement PR. We know Forza will use cloud stuff somehow, so hopefully at E3 we see more about their ideas. I'd really love to see MS make very specific quantitative claims on this, show some data to support said claims, and then have devs show how they are designing for the cloud followed by some in-game footage showing the results.

I think my example of planting C4 on a building strut is a compelling one though. Even if the only mechanism to blow up a building and make it collapse was this kinda time-delayed scenario, that would still likely be compelling for many gamers (assuming the physics simulation is high fidelity).

Maybe some of the technical post mortems on Killzone, Halo etc could be dissected and injected with cloud computing..

I tried posting the KZ demo info earlier in the thread. Also searched for info on BF's destruction system. Here's some KZ info.

KZ:SF Demo:
75MB for animations (600 AI characters, player, any other anims)
6MB for AI
5MB for physics meshes

Here's a BF3 presentation on their physics destruction stuff. You might find it helpful.

http://www.slideshare.net/DICEStudio/siggraph10-arrdestruction-maskinginfrostbite2

 

[-tkf-]

20+Mb/s is he minimum/cheapest available from cable here. In 5-10 years time we'll probably all be on 100Mb/s and up.


I imagine games will stop or throttle background downloads when the best connection is required for cloud stuff, as some apps and games do now. Anything that can be done on the cloud would need a simpler place holder on the local system.

15Mbit tops here, unless i pay a large amount of money. Just a few miles away fiber is an option.

 

[Tap In]

I've just realised another potentially serious factor affecting cloud viability, and that's when downloading content. If you have background downloading active, whether downloading the game you are currently playing or some other game or content, your hardware is going to be saturating the BW. Unless that is throttled, QoS cannot be guaranteed to any level, seems to me. You can't even be sure of 1 Mbps, 30 kilobytes per frame, of content to be available at a given moment. Things like just-in-time calculations as astrograd suggests become even more implausible, unless all of a few kb. Then the cloud service can interrupt any local downloads for the window. It'll of course have no control over bandwidth used elsewhere by the local network. Things like video chat will also impact available BW.

yes there are games on 360 now that even in SP mode, pause background downloads because they are connected all the time (such as EA sports games)


as for rest of house networking yes could be a problem... MS is going to have to have some pretty clear recommendations for this service wrt internet speeds and availability both for devs and consumers

 

[Cjail]

@astrograd

I think you mean "60 AI character" not 600.

 

[3dilettante]

I tried posting the KZ demo info earlier in the thread. Also searched for info on BF's destruction system. Here's some KZ info.

KZ:SF Demo:
75MB for animations (600 AI characters, player, any other anims)
6MB for AI
5MB for physics meshes

That is the allocated area for some fraction of the game's execution time, or their static allocations.
The promised cloud resources for Xbox One by default can contain the whole context of the game, so the static totals don't yield much.

I'm unclear what conclusions we can draw from a storage snapshot, because this may not indicate the actual amount that actually gets used, and because the time component is zero.

 

[-tkf-]

I haven't seen that mentioned. You know the specs of thse CPU's? Link?

Just google data centers, you will find CPU's that was picked based on performance pr watt, no highend parts to be find because of the high costs involved. Cloud strength comes from a lot of parallel power distributed in smaller bits on many servers. The reason most cloud related stuff feels so "fast" like picture services etc is the static nature of the data. Ask Youtube to stabilize a video and it will take 20-30 minutes to "fix" a 6 min long HD video, which is kind a slow but Youtube is stabilizing thousands of videos at the same time which is very fast so clouds are extremely fast at being slow

We had a setup where we had a 40 core cluster working on datasets, this was Nathelm cores. Because the setup was pretty heavy on latency a new setup with 24 AMD core server beat it easily beauce all the data was local on the machine. The raw CPU power was 3 times higher on the 40 Core setup, but latency and cross communication between the jobs ate everything. My point being, that the jobs for a classic cloud would have to be very specific in relation to games. BIG datasets that requires a lot of power but not a low latency response, and for a "fast response low latency cloud gaming service" it would have to be a very special cloud with very dedicated servers/cpu's to a single game.

We know Forza will use cloud stuff somehow, so hopefully at E3 we see more about their ideas. I'd really love to see MS make very specific quantitative claims on this, show some data to support said claims, and then have devs show how they are designing for the cloud followed by some in-game footage showing the results.

Course maker where the cloud runs all the baked stuff would be a good idea. Build your course, pick a time of the day.. submit to cloud, play it the next day, and your friends can download the track as well.

I think my example of planting C4 on a building strut is a compelling one though. Even if the only mechanism to blow up a building and make it collapse was this kinda time-delayed scenario, that would still likely be compelling for many gamers (assuming the physics simulation is high fidelity).

Is a viable but very limited example, unless we make a pure PLANT THE C4 game. But yes, you could build games build around the limitations of cloud response time. How about the toughest Chess game ever!?

 

[astrograd]

And we ask the rest of the house to turn of whatever they are doing because the cloud is busy making calculations?

"Can i watch netflix? NO! i need to play this awesome game that runs in the cloud"
"STOP WATCHING YOUTUBE I NEED TO PLAY"
"Close your browsers, i need to play"

Depends how much bandwidth this stuff actually needs. We still haven't actually arrived at many figures yet, unfortunately. I've tried posting some links and info but nobody seems interested in actually trying to work through these scenarios other than to mindlessly assert they are impossible. One thing though, if the console has background downloads then some of these things can be downloaded and stored locally when you're not playing, no? I don't mean the physics stuff, but things like AI routines and even physics-based npc animations for stuff like hair/cloth/etc. The thick computations can be done in the cloud and sent out to all the players like a patch except devs don't have to hand-make these anims.

Anyone have Skyrim on PC and the construction set? Maybe we could try to use that to help give us ideas on how large animations are in a semi-modern game? I know, Bethesda sucks at animation but still, it's a start!

I'd also like to know about games with in-game cutscenes (basically nothing but physics and animation). Anyone know of any games where that info might be available? How about that Galactic Reign game from SlantSix/MS? Don't they do some battles in the cloud and send back data to have it rendered locally or is it sending back just video frames?