• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Xbox One | Understanding Microsoft's Cloud claims | Tech panel and ArsTech article

I am actually working on a game that uses Azure for procedural generation and the Cloud version is up to 100 times faster in loading levels. (since Azure can generate the procedurally generated data much faster than a mobile phone).

Minecraft for example could generate a world on the cloud in a few seconds instead of 20 or so on 360.

The problem is finding a good fallback scenario for when you have no internet connection. This is the biggest problem atm.

Fuck it man! Online only!
 
Ok, so the article says you can do things like advanced lighting. Yes you don't have to update lighting, and ambient effects every frame... but with the latency.. how many frames WILL it take? How disconnected will those effects be from the actual thing you're rendering? What if you start to turn quickly. You'll have a HUGE disconnect from where something like a real time light shaft would be vs where it SHOULD be.
 
One thing is for sure, if both consoles release, and PS4 games continually end up looking better, Microsoft will find it hard to live down these sorts of claims.

What they are proposing is a logistical and technical nightmare. Potentially considerably complicating the development pipeline further, but that's just my opinion. To me, they would be better suited to just using those servers to offer dedicated servers (which I think they are) for low latency advantages in ping, less lag in online games, bigger maps, more players etc.

Im not an network engineer, but I do feel I have a bit of common sense. Logistically speaking, to take advantage of this stuff as advertised is going to require a skill set that is far beyond from what we are currently seeing in games for something that at best will only gain pairty with late PS4 games. Even if games look slightly better, its gonna be a nightmare to account for a million different setups and connections.

I honestly want to know, in a time where we barely can get worthwhile netcode for a fighting game, where all thes super network planners an engineers that not only account for all the variables, but also orgainse the game and server calculations so they are perfectly in sync are going to come from?

Not only that, but show me the thought process that makes anyone think that graphical flourises that are present in the lead platform for a game that are coded to hardware (PC and PS4) Will easily and cheaply be ripped out and put into the cloud instead? "Oh rip the trees out and put them in the cloud in the xbox version" Am i naive to believe that thats not how game development works?

While you are at it, can you tell me where the publishers are who will be that concerned with graphical parity between consoles that will invest in these people during development, for little no to proof of an increased return because of it?

Finally, can you find me the people who will be so impressed by this that they will buy a game based on it? Nothing suggested here has not been shown to work on a current generation PC and Last time I checked most games on PC dont even come close to tapping that level of power.

This shit sounds like the worse kind of snake oil because while their may be some true to it in practise its not going to be used to anyone bar microsoft and even spordically because of all the problems it can cause if it goes wrong.

The "Cloud" is going to be used many for one thing :- Streaming live video, TV and music to consoles for a fee. anyone who believes different is deluded.
 

Zaptruder

Banned
Can you provide any existing example of a cloud based solution offering this kind of processing?

Nope! I'm just describing what might be theoretically possible with Cloud based processing.

To be sure, I think it's a pretty goddamn complex technical and even economic issue. And that might be enough to torpedo the kind of cloud based processing MS are attempting to allude to here.

But at the same time, I know that I can't say with certainty that they haven't figured this stuff out.

I'll simply take the wait and see approach, while keeping in mind that it's not outside the bounds of possibility for cloud processing to work in real time rendering scene with low latency requirements.
 

Zaptruder

Banned
So the cloud is handicapped to a small number of effects on the periphery. Not that 4x performance boost MS is going on about.

Yes. This is what it seems like to me as well. As developer understanding and skills increase, their ability to exploit cloud processing power might increase.

Certainly, for online and MMO style games, there may be more of a benefit, if we can safely assume that the player will necessarily be connected to play the game (so that in that scenario, less fall over redundancy is needed for non-connectivity moments - i.e. having graceful ways of dealing with no cloud computation).


@phosphor112: If by ambient effects, you mean ambient occlusion - that's a light effect that can be baked into the light map for static objects in the scene, reducing the amount of local processing power required for dynamic objects (i.e. baked into the architecture, vs AO for barrels, crates and other destructable elements).
 

Gotchaye

Member
In your example, pre-computed light maps are too big to store on discs, but not too big to store on the cloud, on the basis that the cloud's stored copy is distributed to many, many clients but the disc copy... isn't? I think I lost your line of reasoning.

The idea is that you don't want to be putting, let's say, 20 extra GB on a disc. 20 extra GB on each and every disc that goes out is a lot of space being taken up, and you've only got something like 30 available (or is it 50?). But that same 20 GB is negligible on a server - you've got terabytes upon terabytes available. At any given time, a bunch of clients are requesting some of that to use in the next few minutes, and you're sending them each MB-sized portions. You don't need to store 20 GB on your server for each client.

I don't have any idea how the bandwidth costs work out.
 
a) if all games need this, then all games need to be always online, 24/7.

b) Cloud cannot fix or help with anything that the hardware cannot handle. The hardware specs still have to be able to handle certain amount of players on screen or certain level of graphics. Unless you are streaming video feed of something you are playing like on Onlive.
 
Man I would love to believe this will make a worthwhile impact. Knowing how each persons connection varies and how bandwidth utilization affects latency makes me skeptical that developers will consider it's benefits worth the headache of making it work. Is it worth putting time and effort into something that an unknown percentage of users will see the benefit of?

I guess time will tell if the internet infrastructure we have is good enough to support this type of solution for most people.
 

Zaptruder

Banned
Man I would love to believe this will make a worthwhile impact. Knowing how each persons connection varies and how bandwidth utilization affects latency makes me skeptical that developers will consider it's benefits worth the headache of making it work. Is it worth putting time and effort into something that an unknown percentage of users will see the benefit of?

I guess time will tell if the internet infrastructure we have is good enough to support this type of solution for most people.

I think it's a direction of future computing, especially useful in squeezing extra performance to keep up with the Joneses over a long generation cycle (this may be the last!).

But it will probably be undersupported and gimmicky till the infrastructure gets good enough and developers and the industry figures out the best way to take advantage of extra high latency compute power.

I mean, I can imagine in 15 years from now, we'll think nothing of World of Warcraft intro quality CG worlds in VR that is been streamed to us via an interactive point cloud, with nothing more than the mobile computing device/smart phone in your pocket. Or even the computer embedded into your HUD!
 

TwiztidElf

Member
I work in a big company. All of our desktop sessions are hosted in our datacenter. There is a massive pipe, way bigger than the pipe from my house to MS (or Sony).
My Outlook and Internet Explorer (for both intranet and internet) response time is absolute dogshit.
If that's the future of gaming, get me the fuck out.
 
If I'm making a single player game would it not be cheaper and easier for me, as a dev or publisher, to just put all the game content on a disk then let the user's console run it all? I can't understand why I would want to spend the extra time, money, and resources to add this cloud component to the game. Besides, wouldn't I need all the technical info, an api of some sort from microsoft, and some training to figure it all out anyway? Just seems like a big costly risk unless I'm building an MMO or something. I mean I could add this to my game for Xbox One and take the risk, or not add it to my game for PS4 and not be bothered with the overhead.
 

supersaw

Member
Until I see side by side video of this in action it's just PR bull for hiding DRM.

Sounds way too similar to the Sim City spin EA was pushing.
 

supersaw

Member
Man 20gb is way, way over what I was imagining. Still, I think I'm understanding where you're coming from, but you're still eventually going to be sending that entire 20 GB to the client, right? Each time they want to play the game through? It feels like a bandwidth deal breaker; not just for the consumer's poor Internet connection but for Microsoft, too. I have a feeling it would be much more cost effective to include a second install disc for that much extra content than to stream it every time to every player.


Not to mention that as a dev what incentive do I have to use this proprietary tech in a game; I will then have to remove it from for my other ports.
 
How exactly are physics not latency sensitive? Is this guy crazy?


There is absolutely no way that this is going to represent itself in any meaningful way in video games. At best you get stuff like faster calculations for asynchronous games/features. (for example, processing AI for turned based games, or simulating games in FIFA/FM, generating random levels, etc).

But there is no way this is going to give you better visuals/details in a game.
 

mrklaw

MrArseFace
Ok, so the article says you can do things like advanced lighting. Yes you don't have to update lighting, and ambient effects every frame... but with the latency.. how many frames WILL it take? How disconnected will those effects be from the actual thing you're rendering? What if you start to turn quickly. You'll have a HUGE disconnect from where something like a real time light shaft would be vs where it SHOULD be.

don't devs already do this? update day/night lighting once a second or so? I've seen racing games and others where you can see the lighting update fairly slowly, but you don't notice because you're moving around.

Maybe that is something that can be offloaded to the cloud, but it isn't something so complex that you'd bother - its more complex to have to write code that can communicate to the cloud, synchronise that with your game, have graceful fallback for network congestion or when the connection is down.

I just don't see many (if any) devs bothering. Decent netcode for multiplayer is difficult enough for some.
 

Trojan X

Banned
Including some little reality to this... If Microsoft want gamers in the UK to have a flawless experience doing this, they better go talk to VIRGIN MEDIA - and others alike - and tell them to stop their stupid "cap the download speed from 100mb to 5-10mb during peak hours" hidden rule. This will badly affect the experience on the Xbone (ah. it's so much easier to say Xbone than Xbox One).

Looks like I will be taking the "wait a see" approach with Microsoft to see how this all pan out.
 
This sounds like it has merit; how does bandwidth play into it? Procedural generation is one situation where I can see a distant cloud potentially being better than a disc of data inside a home console, but it seems to depend on how much data you'll have to transfer to the home console once the cloud procedural generation is done, so much so that I can't see it remaining profitable for Microsoft unless we're talking about really insignificant stuff.

In my case the bandwidth is low but the computational power required is high (as with most procedural stuff actually). This is a very good scenario for cloud processing, not everything is like that.

I have seen some implementations where AI-calculations is spun off async to a cloud server and then interpreted by the client when the call finishes.

I have also seen a demo of somebody (no I am NOT mentioning names), that spins off generating the next level that the player will visit to the cloud while the player is playing, incorporating other players "influences" on the geometry and saving it to the disk in the background. This can also easily be applied to open world games. Updating parts of the city that you are not in yet for example.

It is very much a case of "if you build it, they will come". The more tools I get like this the happier I am as a dev... a pity a lot of the people here just like to moan about it instead of trying to see real world applications and brainstorm about ideas.

It isn't usable for everything, but it will certainly be usable by some new kind of games.
 

amardilo

Member
Yes, which is another thing I don't find to be realistic and potential PR. Especially if the XO sells millions. It just is not financially sound.

It might be fine if they are talking about virtual servers/machines as they can spin up and turn off as many servers as they like (well that their hardware can take). Also it's not likely someone will play a game 24 hours a day so when one console is not using cloud computing the server resource's can be used by someone else.

They can spread the load even more if they can manage resources smartly. There could be 3 virtual machines available to a console and the console can spread the load between those 3 servers (instead of just using 1 at a time) speeding up processing and when 1 server has finished processing a request from another person's console is managed. This could lead to queues and wait times but if MS can handle the traffic effectively it should be OK (i.e. if the load starts to increase and gets to a point before anyone notices slow down more servers are spun up and more of the load is distributed automatically).

Microsoft has spent a lot of time, resources and money over the years on services like Azure (their cloud platform that includes a cloud compute feature to offload CPU cycles) and vitalization (stuff like Hyper-V and their Virtual Machine/Server software and tech) so they should have a lot of knowledge and be capable of getting close to their vision (well maybe not their PR speak).

Also I would imagine to get the most out of this you will need Xbox Live so that will mean MS have a constant stream of money to pay for this.
 

just tray

Banned
Tech sounds promising, stop hating. If you do recall, vgleaks and patents did suggest that the system will be built to take advantage of thus ability and all a game has to do is be designed for cloud and without cloud in mind This will allow Microsoft to update their console hardware and old xb-1 systems will use more of cloud to keep up with the newer zb-1 systems.Think Android to get an idea of what they are doing. Different phones will not run the same apps and games in some cases. They have to be coded to run on tons of different devices.You may have cloud talking to the system while your system talks to a tablet or phone. Regardless of the xb-1 being the new Wii-U around here, I will still buy a One.

Cloud is future proofing your system. It's not imaginary by any means. The One is built for cloud gaming.If it was all about DRM then you wouldn't need 300k servers. I know ten years ago or even at the start if this gen that if someone told me you would be able to speak to your system and it can read your facial expressions I would think its just a dream myself. This is future proofing your system. In 2023 can anyone tell me what new techniques will be available for cloud computing? If anything go off of track record and look at how far Kinect and Xbox live have come?

Can only imagine the possibilities of hacking Kinect 2 on top of cloud computing.I know the science community will love it.

"Better with cloud" I am surprised no one has hopped on that one yet.

Saying everything waiting til after E-3 to decide which system to purchase first. Kind if leaning towards ps4 at the moment.
 

Alx

Member
Sounds like such a dramatic change to development processes that it will never occur outside of exclusives, if at all

It's not too different from the change that developers had to endure when transitioning to multi-core architectures. Their programs used to be sequential and synchronized to the display, now they are threaded and distributed.
 

coldfoot

Banned
So the cloud is this gen's "imagine the possibilities!"? Did that with Kinect and all we got was dancing and fitness games.
 

mr2t

Banned
Tech sounds promising, stop hating. If you do recall, vgleaks and patents did suggest that the system will be built to take advantage of thus ability and all a game has to do is be designed for cloud and without cloud in mind This will allow Microsoft to update their console hardware and old xb-1 systems will use more of cloud to keep up with the newer zb-1 systems.Think Android to get an idea of what they are doing. Different phones will not run the same apps and games in some cases. They have to be coded to run on tons of different devices.You may have cloud talking to the system while your system talks to a tablet or phone. Regardless of the xb-1 being the new Wii-U around here, I will still buy a One.

Cloud is future proofing your system. It's not imaginary by any means. The One is built for cloud gaming.If it was all about DRM then you wouldn't need 300k servers. I know ten years ago or even at the start if this gen that if someone told me you would be able to speak to your system and it can read your facial expressions I would think its just a dream myself. This is future proofing your system. In 2023 can anyone tell me what new techniques will be available for cloud computing? If anything go off of track record and look at how far Kinect and Xbox live have come?

Can only imagine the possibilities of hacking Kinect 2 on top of cloud computing.I know the science community will love it.

"Better with cloud" I am surprised no one has hopped on that one yet.

Saying everything waiting til after E-3 to decide which system to purchase first. Kind if leaning towards ps4 at the moment.

It really isn't promising. People are grasping for straws here. Cloud gaming is terrible.
 

x-Lundz-x

Member
Just like a lot of things, it sounds great on paper.

I know for one I can't wait for me to be playing a game and if there happens to be server lag issues that I get a loading screen every two seconds while the cloud computes my god rays for me.

This is not practical in reality.
 

Sorcerer

Member
Does this mean when you put the disk in the drive that nothing is being installed on the drive from the disk?

Is it just a check and then it installs from MS's servers?
 
Sounds like the same stuff I've been saying in threads, but nobody comments.

What is being described seems to be total BS, and will be generally regarded as such until until an advantage from this tech is actually shown running. Few want to comment on how something so far-fetched could work in theory, except to totally bash it of course.

Especially after hearing very similar claims from Sony in 2006.
 

Cosmozone

Member
Using web services to integrate into games would be very difficult for programmers as has been surely stated multiple times already. And of course people with unstable internet or none are doomed.

Think about the worst experience someone had with the Wii U OS. Multiply it with 1000 and you're not even close.
 

Clear

CliffyB's Cock Holster
The fundamental problem is that if any critical processing is offloaded onto the cloud, your internet connection state becomes the single most important element in the entire pipeline.

You lose that connection... boom. That connection slows or is throttled by local or ISP traffic, how does the system react. Some data (video streams for example) you can adaptively correct by lowering resolution, but not all data is as amenable to that sort of treatment.

Bottom line is of course if this functionality is not available across the board, multi-platform titles wont touch it.
 

andycapps

Member
Assuming that some developers actually use this and that's not all bullshit, how long is Microsoft committing to keep servers up and running for every game? Say next generation after Xbone comes and I want to play the launch Forza for Xbox One, will I be able to? What about the next generation after? I'm talking single player here, not multiplayer. If that game demands an internet connection and a server to build part of the assets used during gameplay, I foresee some type of shutdown eventually where they won't be supporting every game for perpetuity. Again, that's if this is not all PR bullshit, which is much more likely.
 

antic604

Banned
In my case the bandwidth is low but the computational power required is high (as with most procedural stuff actually). This is a very good scenario for cloud processing, not everything is like that.

I have seen some implementations where AI-calculations is spun off async to a cloud server and then interpreted by the client when the call finishes.

I have also seen a demo of somebody (no I am NOT mentioning names), that spins off generating the next level that the player will visit to the cloud while the player is playing, incorporating other players "influences" on the geometry and saving it to the disk in the background. This can also easily be applied to open world games. Updating parts of the city that you are not in yet for example.

It is very much a case of "if you build it, they will come". The more tools I get like this the happier I am as a dev... a pity a lot of the people here just like to moan about it instead of trying to see real world applications and brainstorm about ideas.

It isn't usable for everything, but it will certainly be usable by some new kind of games.

Fully agree with you here - the whole idea sounds like there's a lot of potential there! The real question for me however is is it platform dependent? I mean it is a software thing on the platform side and what you only need is a server you'd send your request to, right? If so, what sort of advantage will Xbox have over PS4? I understand MSFT will maintain 300k servers to facilitate the 'computation requests', but will it do so for 3rd party developers as well? And who'd pay for that - XboxLive membership? If so, what's stopping Sony to do the same? Hell, they already probably have a lot of Gaikai servers they could use?
 

oVerde

Banned
a) if all games need this, then all games need to be always online, 24/7.

b) Cloud cannot fix or help with anything that the hardware cannot handle. The hardware specs still have to be able to handle certain amount of players on screen or certain level of graphics. Unless you are streaming video feed of something you are playing like on Onlive.
1) no. Games will provide fallback solutions, but in some years the diferencea of 1.3Tf to ~4Tf will be drastic.

2) no. The calculation will be done it is so much light to the hardware is almost just add to video buffer and show it next frame.
 

SPDIF

Member
As someone who has been working with Azure for years, I had a pretty strong reaction to the PR statements. Just to clarify, I do not work for MS nor in gaming; my work is business app related - a field where MS has been pushing it's cloud offering for a while now.

First of all: These servers have been around for a while, I doubt these 300k claims are based on something Xbox specific. If someone knows better please correct me, but I believe 300k is just the total number of all MS cloud servers currently. So for example the server running our company email is already listed.

Calculating power: The current servers are no graphical beasts, they have 8 core processors and a lot of memory, but quite basic GPU:s that are shared with up to 8 other customers ( Everything is virtualized ). Currently the servers mostly host databases, websites, sharepoint deployments etc. While they certanly might be 4x Xb one in CPU calculations, I doubt they will be doing any heavy graphics calculations anytime soon.

Price: A singe 8 core machine runs at about 500 dollars a month at the current pricing model. I just don't see how it can be feasable in the long run in games that have no monthly fees. If they actually want to calculate anything worthwhile graphics related, we are looking at a several cores per user. It simply makes zero sense to build complicated cloud support for just one extra core.

What the Cloud is good at: The cloud is good for handling large amount of data ( Disk and memory is cheap ) and doing userbase wide calculations. It might calculate statistics or auto balance the game based on a lot of user data, or it might store persistent world states for large worlds ( Online destructible environments ). But actually improving graphics, that will most likely not happen; they would lose their game profit way to fast. Again, 8 cores cost 72 dollars for 100 hours of usage at current Azure prices. How much are those shadow calculations worth?

What we will see: EA will propably make websites for all their sport games based on in game statistics and data. Lineups will be handled trough the cloud and you can compete with your dream teams or whatever. I believe most of the "exclusive" EA sport features are related to this. The big advantage in hte cloud is accessibilty and data. I see more companion iPad apps like Destiny has in the future. You can most likely see Cod statistics live from any device.

Why I am sceptical: The cloud is nothing new and Azure has been around for years. How many games use cloud calculations currently? I only see this as a way for MS to bring current web based features under one roof / API. Thus giving MS a steady revenue, and simplifying development of these kind of "off-console" experiences.

Nope. Xbox Live has always run on its own servers, they've never previously run on Azure. And these 300,000 servers that Xbox One will be running on will be brand new servers dedicated to the X1.
 

sajj316

Member
All this cloud computing talk is taking me back to PS3 and "distributed computing" where the only application that realized this vision (that I know of) is Folding @ Home. Theoretically possible but I wouldn't hold my breath
 
Fully agree with you here - the whole idea sounds like there's a lot of potential there! The real question for me however is is it platform dependent? I mean it is a software thing on the platform side and what you only need is a server you'd send your request to, right? If so, what sort of advantage will Xbox have over PS4? I understand MSFT will maintain 300k servers to facilitate the 'computation requests', but will it do so for 3rd party developers as well? And who'd pay for that - XboxLive membership? If so, what's stopping Sony to do the same? Hell, they already probably have a lot of Gaikai servers they could use?

Gaikai is totally different. It isn't a cloud application platform.

Anyway... yes it will cost MS to run these servers. I guess they will pay for them. I cannot see how else they will offer devs a reason to use them.

Sony doesn't have a real cloud platform, but yes they can use existing ones by also just paying for it (they could even use Azure as well). The strength of MS will be tight integration with Live, X1 and Azure. That will make it easier to include this on X1.
 

ASIS

Member
Wait, I don't understand. If cloud is so important to the power of the system how can you play offline then? Sorry if this had been already discussed.
 

ymmv

Banned
The fundamental problem is that if any critical processing is offloaded onto the cloud, your internet connection state becomes the single most important element in the entire pipeline.

You lose that connection... boom. That connection slows or is throttled by local or ISP traffic, how does the system react. Some data (video streams for example) you can adaptively correct by lowering resolution, but not all data is as amenable to that sort of treatment.

Bottom line is of course if this functionality is not available across the board, multi-platform titles wont touch it.

Programmers have to make the game function in both online and offline mode. You can't have the game suddenly display a pause mode because your internet connection has a hiccup. It will also make the development of games that rely on cloud processing more complex.
 

LukeTim

Member
I am sceptical of this cloud thing...

I will wait to see it in action to pass judgement, hopefully it will be at E3... however, I would have thought that for video-games pretty much everything has to be computed for each frame, so I wouldn't have thought there were any latency insensitive computations which were important to the gameplay...
 
I can't imagine that MS invested in these resources for nothing, and I think we'll actually see some amazing, "Always on" exclusives that take advantage of it.

It sounds like too much work for multiplats unless the one dominates the market. We'll see.
 
It's not too different from the change that developers had to endure when transitioning to multi-core architectures. Their programs used to be sequential and synchronized to the display, now they are threaded and distributed.

Difference being

PC, PS3, 360 all went multicore multithread development. In case of cloud computing

Xbox One is the only platform going for it

PS4 and PC...won't so this will be used by first party quite a bit. But as I stated earlier in the thread, third party developers will go the path of least resistance. The only way this is going to take off in a major way is if MS dominates the marketshare.
 
Top Bottom