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

[Shifty Geezer]

Well that's a bit harsh, but what about you give examples of things you know/expect to work when being moved server side ?

The only one I know for sure could work is AI in some RTS games since it might be run at low frequency, what else do you have in mind ?
(In details please, so we can construct our own mental images of the process and its consequences.)

Xenio is not blessed with clear communication of this ideas in English, which is where a lot of these discussions fall down. I don't know if he has anything more than a general idea either. But those of us with a better understanding can certainly explore what's possible if we factor in time-critical aspects.

Let's break a game down into known components.

User interaction - time critical
Graphics rendering - time critical
Dynamic lighting - time critical
Game physics - time critical
Environment physics - needs to be realtime but maybe some tolerance allowed?
AI - can tolerate some delay (we need to put a value to this)
Environment AI - persistant worlds, non time critical
Environmental transformation (plants growing, buildings ageing, people ageing) - non time critical
Something new we haven't thought of yet?

Within the scope of games, certainly certain genres, I'm seeing opportunities to provide a greater experience than just a console with a BRD drive can support by using online servers, either for data or computation. That's the basis of MMOs, so we know it's possible.

 

[Shifty Geezer]

This seems more realistic to me, but how much bandwidth would such a solution use and how often would it need to be updated? assuming 1080P fidelity would such a data transfer be rather large?.

Possibly. But it may not even be data like 2D maps. Some GI lighting can be very low fidelity but still look right, because GI is typically soft. So you may only need a few thousand samples per viewpoint, and a few MBs data for a region. I don't really know. Anyone well versed in lighting solutions may have a better idea. Would be nice if Graham turned up with his experience of Geomerics (if I've got that right!).

What you question is basically the major limiting factor though. For non-time-critical calculations, the limiting factor will be BW, and possibly data caps. A target of 1 MB/s seems realistic to me, so the data would have to limit itself to that and structure itself to those limits. Game design would shift to using more computation resources (if these servers are even economical) and more compact data representations.

 

[(((interference)))]

I posted this in the distributed computing thread, but I think that at some point in time, Xbox One's total power was comprised of raw computational power + cloud. So Total Power = 1.2TF + Cloud. This would result into an always online system. The leaked VGLeaks documents did hint at an always online console. I think that due to recent backlash Microsoft had to get rid of that idea, leaving us with a 1.2TF base model.

What do you guys think?

Just....no

...........

With the 200GB/s bandwidth figure lets not forget that last gen MS were claiming the 360 was packing 1TF (with Sony claiming 2TF).

So I'm betting its a similar case of creative accounting we're seeing here.

 

[Grall]

There is certainly scope for, instead of rendering these offline and saving precomputed data on the game disk, rendering them online and downloading the results. This way you could support dynamic day cycles with full environmental GI, with lighting calculated on servers way ahead of the immediate frame requirements and cached on the consoles ready for use in several minutes time.

Yes, but if the data is precomputed anyway what would be the gain versus just loading said data from local storage? I see none, only drawbacks. For one, you'll be reliant on 1: server farm being online (and not DDOS'd, etc, suffering network equipment failure, ISP issues and so on), 2: your own internet connection being up and available. And a server farm costs a lot of money building, and potentially even more running, with power, cooling and other utilities, staff, rent and bandwidth costs quickly piling up.

Rather than these far-reaching, Ken Kutaragi-like pie-in-sky ambitions that are fragile and costly in nature, I think MS's cloud strategy is more like everyone else's cloud strategy in reality, which is mostly stuff like cloud data storage and synching, and running multiplayer servers, including MMO game logic.

If pre-calculating GI is tough work, pre-calculating GI for tens of thousands of players is going to be tougher still. Even if you pre-calculate ahead of time, you'll still have to be able to stay ahead of the player, and with many players loading the servers that becomes a major issue I would think. On the other side of the coin, a BR disc is big, you can fit a lot of GI data there I wager.

 

[Cjail]

@Xenio

I see several problems

First the main problem is internet.
Either latency is too high or upstream bandwidth is too low.

Second Cloud is an order or magnitude more powerful than console so needing/relying on the console console for certain task at all effect means handicapping it.

Third cosnle offer far better performance than cloud ATM so cloud as well can also be an handicap.

 

[Shifty Geezer]

Yes, but if the data is precomputed anyway what would be the gain versus just loading said data from local storage?

Available data. You can fit 50 GB on a disk. You can stream the entire world from Google Maps. But it's not a strategy I particularly advocate. Just seems more likely to me use the data aspect of servers than the processing.

I see none, only drawbacks. For one, you'll be reliant on 1: server farm being online (and not DDOS'd, etc, suffering network equipment failure, ISP issues and so on), 2: your own internet connection being up and available. And a server farm costs a lot of money building, and potentially even more running, with power, cooling and other utilities, staff, rent and bandwidth costs quickly piling up.

Yeah, it definitely has operational concerns, effectively making the solo player experience something of an MMO experience. Always connected, always dependent on another parameter. The local game would have to have some backup, so when the service is down the game can still be played with none of the fancy extras. At which point is it worth investing in? A few games, like living world RPGs and city roamers, would probably benefit, but then that's more of a game server and you'd go multiplayer with that, I'm sure. More chance to monetise.

Rather than these far-reaching, Ken Kutaragi-like pie-in-sky ambitions that are fragile and costly in nature, I think MS's cloud strategy is more like everyone else's cloud strategy in reality, which is mostly stuff like cloud data storage and synching, and running multiplayer servers, including MMO game logic.

TBH I haven't heard MS's specific description so I have no interpretation of their intentions. But the discussion itself as a platform-agnostic concept is still worth exploring. If we all expect games to go cloud anywhere, is there an interim between local games and streamed video?

If pre-calculating GI is tough work, pre-calculating GI for tens of thousands of players is going to be tougher still. Even if you pre-calculate ahead of time, you'll still have to be able to stay ahead of the player, and with many players loading the servers that becomes a major issue I would think. On the other side of the coin, a BR disc is big, you can fit a lot of GI data there I wager.

I would expect, to mitigate costs effectively, the living world simulation would operate for hundreds of players at least per server. So basically more MMO than cloud computing for local experience. But that wouldn't prohibit a solo game using the same server computations. Let's take GTA6. There's a city with a populace of 100k AI bots and 1000 real players. There's lots of environmental interaction, destruction, building, whatever. The server would calculate the background assets that the local console assembles into the rendered image. Those same assets could be used by a solo player living in the same city, seeing it change around him.

The economics of whether distributed computing TBH don't really interest me. At the moment I'm more curious about what is technically possible and how servers could be used to expand the local box experience.

 

[Daozang]

AI - can tolerate some delay (we need to put a value to this)
Environment AI - persistant worlds, non time critical
Environmental transformation (plants growing, buildings ageing, people ageing) - non time critical
Something new we haven't thought of yet?

Within the scope of games, certainly certain genres, I'm seeing opportunities to provide a greater experience than just a console with a BRD drive can support by using online servers, either for data or computation. That's the basis of MMOs, so we know it's possible.

I think that the reason we haven't seen these things yet in games, have to do with production cost, and game design.
It seems to me that processing power comes at the end of a long list of reasons.
I also believe, that it's because of the same reasons we don't have rpgs as deep as Planescape Torment or as expansive as Daggerfall.

 

[babybumb]

This seemed even more "out there" for any meaningful performance gains for single player games. At some point it just becomes easier to render the whole game on servers when they are much more powerful than the console anyway which is the way Gaikai is going

 

[Scott_Arm]

Data centers are massively complicated and expensive. Giving everyone Xbox One or PS4 level processing out of the cloud would be absolutely insane. Giving them a small amount of additional processing is possible. It's an interesting debate to discover what type of computing is most reasonable. I imagine it'll be something like a job system, where games can leverage an API that sends data and a "program" to the cloud for computation. Something like GPGPU model. Dedicating cores to each user would not make a lot of sense.

They could do interesting things with group AI. Individual behaviour may be too time sensitive. Think of a game like GTA where you might have factions/gangs or groups of people, and you need to make group decision making. You tend to see very simple group behaviour in games. They like you or they don't. They do not exhibit any kind of leadership structure or group action. That's one area where the cloud could help significantly. You could have hierarchical structures with roles. Skyrim is another good example with warring factions that really never seem to be at war. They just wander around a bit and fight when they pass each other randomly. There isn't any kind of sophistication in how they think as a group. Just a thought that came to mind.

 

[3dilettante]

Some of the things I'd imagine being doable early on are tasks that are typically awkwardly tossed back and forth between clients in multiplayer games anyway.

For an FPS, it's things like final arbitration of hits, movement, and level state in a match.
These are frequently affected by lag anyway and involve trying to reconcile multiple asynchronous systems that are supposedly working together on the same world state.

Could this be done by a publisher's own servers, or say we were playing a decade ago--our own dedicated servers?
For the former, it's less work in a field many publishers have shown they have very little skill with, and for the latter--this isn't about increasing player freedom.

The Microsoft service should take away some of the motivation for companies like EA to kill their online servers so regularly. Unless the cloud code is on the client discs, which it doesn't need to be, the publisher still maintains an anti-piracy measure and the option to kill the service when it is most profitable to force people to upgrade.
There are probably validation processes for the cloud clients like there are for the client-side code, or there better be.

The level of computational offload and memory consumption is going to be controlled, at least early on. Large game releases can have millions of copies active at the same time, and they can't all be capable of monopolizing a cloud server.


I'd wish there were an option to roll my own server, in part to avoid my games falling apart when Diablo X comes out, to allow me to play my games after the publisher poison-pills their cloud client, and to serve as the cloud version of a UPS (to answer those who use the "power goes out, too" argument).

 

[patsu]

What's wrong with MMO ?

Progressive download will be handy.

Game/app streaming a la OnLive and Gaikai can be mixed in with local games too. Like having your own remote VM on the server.

 

[ERP]

I think the bigger issue is going to be convincing anyone doing a cross platform title to use it at all.
Any use of a remote resource like this is going to take significant planning, plus implementations in multiple environments.
Given to do the same on the competitors platform I have to host the servers, why would I go to the effort?
There is also cost, Live as it is right now certainly won't pay for the cloud if it is broadly used, so the cost probably falls on the publishers, another reason not to use it.

 

[patsu]

Yeah... IBM's Butterfly effort tanked because the server needs you to program the game in certain way. They had (have) no relationship to jumpstart the initiative.

If MS already have the developer base writing to their API, then perhaps the onus is on them to make their middleware work across the machine boundary.

Still, from user perspective, how is this different from an MMO ?

 

[Cjail]

@ERP

Cloud it's not going to be free for the customer because not even internet is free on Xbox but if it's not popular costs will fall on the publishers.
And even if they choose to make it free, to popularize it, then costs will fall entirely on the publisher's which would be a big disincentive.
Also what level of freedom will publishers have?
If the freedom given to the publisher is somehow limited then cloud becomes even less appealing.

CCP choose PS3 because XBL polices are too restrictive while Sony allowed them to do pretty much what they wanted on the PSN with little restrictions which tells me that MS is not very open to others.

 

[patsu]

Microsoft will have a deciding advantage here because of their business cloud division. They are already making money from the infrastructure and can expand aggressively.

The payment publishers foot should be lower than rolling their own.

 

[ERP]

Microsoft will have a deciding advantage here because of their business cloud division. They are already making money from the infrastructure and can expand aggressively.

The payment publishers foot should be lower than rolling their own.

There are internal projects at MS that use EC3 instead of Azure because of the cost.
And if a publisher rolls their own they can use it on both platforms.

 

[patsu]

For large or small scale projects ? Do you mean deploying your own Azure servers, or renting Azure virtual servers ?

 

[ERP]

For large or small scale projects ? Do you mean deploying your own Azure servers, or renting Azure virtual servers ?

At least when I was at MS, Internally you bought time on the Azure servers just like external customers. You can't deploy your own Azure servers AFAIK.

Azure is not a cheap cloud service, and I'm wondering who's paying for it.
It's OK to talk about leveraging the cloud, but someone is paying for the machines, now maybe MS is subsidizing the cost, but live subscriptions as they are today would not cover extensive cloud usage by games.

 

[Gubbi]

At least when I was at MS, Internally you bought time on the Azure servers just like external customers.

At full price? That's insane.

Azure is not a cheap cloud service, and I'm wondering who's paying for it.

iCloud is running on both Azure and Amazon elastic cloud, so that is at least one customer.

Depending on usage pattern Amazon varies from half price to slightly more than Azure. The higher and more constant you demand is, the lower the hourly price.

It's OK to talk about leveraging the cloud, but someone is paying for the machines, now maybe MS is subsidizing the cost, but live subscriptions as they are today would not cover extensive cloud usage by games.

Users only play one game at a time, so while the demand varies from game to game, the overall demand on cloud servers should be propotional to the number of users (ie. fixed), since Live revenue is proportional to the number of users as well I can think of a working business model for it.

That said, beyond dedicated MP servers, I think it is a load of bull. At least at this moment in time, maybe in five to six years.

Cheers

 

[patsu]

It is possible that MS want to balance general Azure rental vs buying your own licenses to maximize their profit. Azure customers should be larger enterprise customers who are heavily invested in Win tech.

If they see cloud computing as an inherent part of their media future, I doubt they will overcharge. Their reference price will be publishers rolling their own on EC2 (or Google's farm). There are also XBL customer service and analytics backends that they can cover for a small premium.

It's all their own software. The infrastructure cost is amortized across all the large Internet sites they own. They are making money from their business cloud. Most of the publishers won't have those advantages even if they could rent from Amazon and Google.

Their main challenge would be if Google come in with their own gaming specific cloud.