The cloud is just over the horizon [2018-2019]

fundamentally, for this to work, both processing power and latency on the Azure side to the customers have to hit a specific minimum, otherwise certain groups of players will be unable to make this work. This is where we need to ask if the global build out was inadequate enough to even support this service as is, and they've been making adjustments to the platform since.

On the software side of cloud computing and the client as well I'd agree with the goal of minimizing the amount of data transfer for this to make it all work.
 
Shifty Geezer said:
With Crackdown 3, we focused on the hardest problem first: the distribution of a very complex physics simulation.
Click to expand...

We're back to the question of what gaming compute problem does this solve, followed by: could it be better solved be pre-calculated / near approximated calculated rather than doing it realtime. A lot of modern algorithms solve [in realtime] what can be indistinguishably solved by clever fakery.

I think Microsoft's approach is 100% the future of gaming but right now? My question comes back to the only gaming question that matters. What difference does it make?, will anybody notice?, does it make the game better?
 
  • Like
Reactions: JPT
DSoup said:
We're back to the question of what gaming compute problem does this solve, followed by: could it be better solved be pre-calculated / near approximated calculated rather than doing it realtime. A lot of modern algorithms solve [in realtime] what can be indistinguishably solved by clever fakery.

I think Microsoft's approach is 100% the future of gaming but right now? My question comes back to the only gaming question that matters. What difference does it make?, will anybody notice?, does it make the game better?
Click to expand...
You mean approximated by machine learning models ? If so I agree for that being an effective local route. Not sure how if it works without seeing it in action.
 
iroboto said:
You mean approximated by machine learning models ? If so I agree for that being an effective local route. Not sure how if it works without seeing it in action.
Click to expand...
Not ML specifically, any kind of fakery. The kind of fakery that video game devs have employed for nigh on forty years.
 
DSoup said:
We're back to the question of what gaming compute problem does this solve, followed by: could it be better solved be pre-calculated / near approximated calculated rather than doing it realtime. A lot of modern algorithms solve [in realtime] what can be indistinguishably solved by clever fakery.

I think Microsoft's approach is 100% the future of gaming but right now? My question comes back to the only gaming question that matters. What difference does it make?, will anybody notice?, does it make the game better?
Click to expand...

I agree, but the part that will be the roughest, I guess. Is the 60 FPS fighting games, I never played any of them online, but from what I read, FPS/low latency is what makes or breaks those games. In general anything that is 60 FPS and how about VR? :)
 
It's hard to believe that the contracts MS had with Cloudgine was not protecting the project from a buyout.
We are talking about one of the biggest company on the planet and while they may do many stupid things from a consumer pov I am sure that their legel teams are top notch.

In my opinion is that the whole cloud processing thing doesn't make any sense from a commercial point of view.
It was mostly a marketing stunt to promote Azure and led people to believe that Xbox One was not that underpowered.

Cloud destruction is expensive, unreliable and doesn't yields many benefits.

Cloud processing make sense with gaas products and even in those cases, the less money is spent on it the better.

Whenever the processing is moved from the customer the question that has to be asked and answered is: who is going to pay for it and why?
 
lefantome said:
It's hard to believe that the contracts MS had with Cloudgine was not protecting the project from a buyout. We are talking about one of the biggest company on the planet and while they may do many stupid things from a consumer pov I am sure that their legel teams are top notch.
Click to expand...

From the outside Microsoft's relationship with Cloudgine appeared to be one of a licensee of Cloudgine's propriety technology, which does not protect Microsoft from the consequence of any other company acquiring Cloudgine. Of course under contract law Epic not only bought Cloudgine's technology but also their liabilities, including ongoing contractual agreements like licensing.

This acquisition is nowhere near as significant as when EA bought Criterion Games and, inter alia, acquired the RenderWare engine which at the time was powering the GTA III era Grand Theft Auto games. You know another game that used RenderWare? Crackdown. :runaway:

They must be feeling a sense of déjà vu over at Microsoft.
 
Tkumpathenurpahl said:
At this rate, it's going to be the other way around.
Click to expand...

What rate are you talking about? Do you have knowledge that Halo Infinity releases before end of January 2019? Crackdown releases in February 2019. Regardless of it's cloud interaction.
 
lefantome said:
So was it all PR?

Looks like it.
Click to expand...
It didn’t make a large impact on changing how the game is played. In that sense yes it’s PR.

In that they actually accomplished it, and it works by syncing all clients to show the same damage using the same vectors, yes they succeeded.

They are as far as I know, the only game company right now to have this under their belt. Cloudgine is the only other studio with this type of experience.

If they can find a proper use for this technology that would be ideal for their investment in this area.
 
Viva Pinata Battle Royale

iroboto said:
In that they actually accomplished it, and it works by syncing all clients to show the same damage using the same vectors, yes they succeeded.
Click to expand...
CD3 MP currently can't even have friends playing in the same party/session because everyone can be around the planet at variable distances to a particular game session, so that's kind of a party pooper if it ultimately means A) a variable online experience just to keep folks together or B) limiting the party to yourself.

This just seems like a bad end-user experience with no great solution because it's inherently flawed unless you have no friends and typically play dvd movie experiences.
 
Last edited:
iroboto said:
In that they actually accomplished it, and it works by syncing all clients to show the same damage using the same vectors, yes they succeeded.
Click to expand...
How is that different to other games with destruction? From what I've seen, debris in Crackdown 3 is cosmetic. You don't climb on it and it fades away after a bit. Is'nt that how other games like Battlefield handle destruction?
 
MrFox said:
Maybe they will figure out they should redirect their efforts on a PSNow competitor instead of trying to offload what doesn't need to be offloaded...
Click to expand...
Ms are working on that, having recently showcased a new streaming tech concept (delta graphics streaming).
 
Shifty Geezer said:
Ms are working on that, having recently showcased a new streaming tech concept (delta graphics streaming).
Click to expand...
Unless I missed something that's still augmentation of the local game. It needs to download the game, run it locally, and then enhance the visuals rendering the game offsite too.

To get the real streaming convenience (versus the other distribution models, or augmentation) it needs to stream games like netflix streams videos. Click on any of the hundreds of games, and it starts immediately. Maybe their next service is implied to also work that way though... If they can augment the local game, they kind of have the infrastructure to stream without any local copy. Like a choice depending on your connection quality.
 
You must log in or register to reply here.
Back
Top