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

I understand but that ship sailed for me when I subscribed to Office 365.


so hard part is over for me.. if the internet becomes unusable and renders these devices useless for us at some point or MS becomes an asterisk in time, we probably have bigger things to worry about at that point.


As for ownership, I never had a problem "licensing" someone's IP, paying them for it while it was available, it's a service rather than a product in my mind.

 

Its not big yet but it will

 

Would I be right in assuming that simulating something like a flag blowing in the wind would require a small set of data, but be more bandwidth intensive? You could push a small amount of data to the could, let the cloud deal with the bandwidth intensive computation, receive a small set of data s the result?

 

So by now (with current gen) they would have 60 million times 3 the power of XBOX 360 in the cloud. Someone do a calculation on how much CPU power that would require. Not to mention the power bill.

 

There are now 8 of these data centers with two coming to Australia and one to China. There may be others. There are also at least 24 content delivery points on top of that.

http://en.wikipedia.org/wiki/Windows_Azure#Datacenters

That's a lot of data center, but it's obviously doing a lot more than just Xbox Live.

 

So they're only looking at 100,000 Xbox's sold then? Wow, talk about putting up an easy target to hit........guess Win8 made them a little skiddish.

 

No because with any "cloud" infrastructure you don't assume everyone is online simultaneously.
MS has several different uses for it's data centers.
Bing has a lot of machines, and they use their own infrastructure.
The Azure infrastructure is probably what XBOne is sharing. So at least some of the infrastructure likely already existed.

 

That does put a damper on the expectations some have for certain kinds of persisent worlds, or simulations that can run while the machine is off.

A popular title like that could leave a much higher number of cloud instances running than would be assumed otherwise.

It also puts a damper on the pie in the sky idea that such games would be able to reserve a whole high-end server farm to themselves, but such is life.

 

Well that depends on how big the flag is........but yea thats certainly an example. You might be talking the color values of a hundred pixels, chicken feed in the bandwidth picture but to calculate what the proper values are may go beyond the available local processing capacity.

 

Ok, that's helpful to know. I guess another way to look at it would be to take a game like Borderlands 2 (maybe?) on the PC, that has a lot of animation of the environment and see how much processing that takes up on the CPU/GPU, and then figure out if how much bandwidth you'd need to the cloud to make that possible.

 

Azure must be doing something else or Microsoft is having a hard time selling their cloud services, it's not like a endless pool of power.

If they really believe in this and we are to believe that they will supply 3xtimes the power pr console then there will be a demand for lots of CPU power. Some of it can be handled by the different time zones (depending on lag vs geographic position of server centers). But if developers pick up the torch and build their games with Cloud in mind and uses this extra power, then it's fair to assume that there will be a significant number of servers needed to be online serving that demand. And even if it's only 1 million gamers that play at the same time, be it single or multiplayer that would be 3 million times the XBOX One CPU power at least. Considering that halo had 500.000 people playing just that game pr day, online i would guess that 1 million to be low. And Microsoft would have to plan for giant peaks during holidays, new game releases etc. They can't have customers seeing ugly graphics during peak hours...

 

That's why I keep saying someone has to be paying for it in the end, and I don't think $60 a year live subscription is going to come close to covering it.
For the persistent stuff, you have to charge the player, much like MMO's.

 

I don't buy it.

For all the hoopla around here about gamers having unrealistic expectations in wanting something greater than a 120mm2 1.2TF GPU in xbone, we are now suppose to believe that for every box sold, they have invested in a 3.6TB server? Yeah... don't think so.

 

Unless they're referring to CPU power and not combined CPU and GPU. Either way, it sounds like a very tall order. More of PR than anything.

That's why I think it's probably better to find info on a particular game and see more much time they spend processing physics that would be "suitable" for the cloud. Then you might get a good idea of things that could be offloaded and how much processing that would require on the cloud end. Probably the best way to discuss the feasibility of this idea.

 

I don't doubt that there are efficiencies to be gained if these same servers are used 8-5pm in work environments and then after 5pm are used for gaming environments, but this is still unrealistic for any kind of volume.

This screams, "don't write off our weak spec machine! Cloudz!"

It would have been better to just suck it up and put a proper apu in the box, or worst case scenario, scrap the design and source off the shelf CPU + GPU.

Much less costly, and real tangible benefits right of the bat.

At that point, tapping into the cloud isn't a necessity, but a luxury.

 

I wasn't able to find the specs on the servers MS are using, other then that they are AMD powered. That being said, a single server can have anywhere between 50 - 100 CPUs and GPUs. If they're using AMD CPUs, they're likely 16 core Opteron's, if the GPUs are AMDs that's anywhere between 1.3 to 4 TF per gpu for the most recent firepro line. A single midrange (est 75 cpu/gpu) server would have the power of about 130 - 150 xbox Ones. If there are 300,000 servers, and MS said each dev has access to 3x the power of an X1, then MS has capacity for about 13 million consoles at launch by my rough math.

 

I guess people are reading different articles than me, in reference to the cloud, they specifically stated 3x CPU and storage. No mention of GPU on the servers.
That's still a lot of general purpose computational power.

I would expect that a game using the cloud would have cloud side data (a cloud install) which may or may not differ from the game. Most of that data doesn't need to exist per user, so it can be really big potentially.

I'd like to see the experiments they've done, and I think it would be interesting to play with to see what you can really practically do with it. I wouldn't want to do those experiments while trying to ship a game, so I'd expect MS to provide at least some components that are relatively easy to integrate.

I would still like to understand who is paying for it. It could well be that there is a surplus of azure compute available at peak gaming hours, but I can't see that being enough.

 

That's a good point. Some MS ppl just said it was 3 times the power of X1. Did you have a link showing them specifying that it's CPU power they use in that loose language's comparison?

 

The largest AMD servers only support 4 sockets; the fastest configuration giving you 64 CPU cores at 2.5 GHz. That's only 12.5X more total Gigahertz than a single XBONE. Assuming that Jaguar is somewhat worse per clock than the Opterons, at best a single server has the total throughput of 15 XBONEs.

And I have my doubts that Microsoft are using 4 socket servers, from what I can tell most Azure servers are using two socket K10 based Opterons with only 4 cores.