Xbox One: The Cloud Could Save Gaming

In The Gaming Community, Video Games on June 20, 2013 at 2:31 pm

Like most of Microsoft’s PR about its new console, the alleged 300,000 server-powered Cloud behind the Xbox One is ill-defined and subject to a great deal of controversy, labelled as anything from an always-online DRM excuse for single-player games to “a lie” as spun by Witness dev Jonathan Blow.

While what it is has yet to be fully settled, what it could be is the salvation of the next generation of gaming.

Despite being one of the more intriguing features of the Xbox One, details on how its cloud will work, how it will benefit gamers or how it will benefit developers have been surprisingly sparse.

In all of E3, there was a single presentation that focused on this feature, and even that was far from in-depth, focusing more on ideas rather than specific cases.

Despite this, what was shown is very promising.

Examples were given about offloading intensive processes to this server cloud, allowing for computations that would both be removed from the Xbox One’s hardware load as well as not limited by its current capabilities, offering returns far greater than what could be achieved through the box alone.

Described was a simulation of our solar system with hundreds of celestial bodies — planets, moons, comets and asteroids — being modeled and tracked via the cloud. What would normally be a fairly static offering is instead dynamic and fully realized as the cloud takes the brunt of the calculations and merely offloads the finished product to the console.

Other examples given are in AIs that can begin to approach something human in their routines: complex, dynamic, unpredictable, and since its in the cloud, not stressful on the console producing it. Or that of an open-world game that truly has persistence, with much of the simulation offloaded to the cloud.

Imagine Skyrim where the world didn’t start and stop depending upon your cone of vision. You can see the limitations in gameplay; the AI distance is shorter than the draw distance, so far-off but visible NPCs are reduced to simple actions until you get close.

There were discussions of latency, but when you’re dealing with rounding out a dynamic world, weather patterns and foliage renders hardly need to have a 0/ms lag time.

These are all very exciting ideas. They do, however, sound very much like every other ill-defined feature that gets a few examples tossed around but then never fully utilized.

I also draw an internal comparison to the PS3’s complex cell architecture. Powerful, complex and unique hardware features have a habit of getting left behind by all but the most dedicated first-party developers.

Which is why Microsoft needs to do the following about the cloud, if it isn’t doing so already.

The cloud should be more than an optional place to offload intensive processes. It should instead be a storehouse of APIs and library functions that developers can easily call to for their games.

So for your open-world game, instead of needing to license SpeedTree and Havok, you would just use the Xbox One Cloud APIs for foliage and physics. Ideally, these would integrate well because such features already exist as piecemeal functions inserted and then called upon later, so it wouldn’t be so intensive or far outside what they normally do.

Which would mean your normal multiplatform title would run better on the Xbox One because tons of intensive processes are not weighing the console down, and are being rendered in higher fidelity via the cloud processing anyway.

If Microsoft strove to provide a truly comprehensive library of such functions, development costs would be reduced for Xbox One titles as devs wouldn’t need to spend inordinate amounts of time dealing with producing shaders, AI, world sim calculations and more.

It’s a very exciting prospect, and one that obliterates any notion of how the RAM difference between the PS4 and Xbox One might play out. It’s something I’d be fully willing to deal with an always-on connection for.

  1. I think the promises of what cloud computing could do are at least interesting, but I have yet to see a game where I’ve had a want for what it has to offer.

    I guess I would have to start by asking the greater purpose in running the game world beyond the player’s cone of vision, if the world experience itself is going to remain the same. If that space beyond a mountain doesn’t exist until I cross it, how is the game experience better served by it existing anyways when I’m so far away from it?

    If our systems are going to get more powerful, does it become necessary to have an outside line that can do all these things if the local device can handle it well enough on its own? Or are we operating on the assumption that optimal graphics will always bite off more than the rest of the system’s processes can chew?

    I do concur that the presence of cloud processing would (and theoretically does) greatly boost what the X-Box One can do over the PS4, but it does come with the truth that it would have asked much more of players than the already laborious 24-hour connection check would have required.

    For what Microsoft was proposing, cloud processing already seemed (and even more so now) like an accessory rather than a built-in feature, much in the way the Kinect and the PlayStation Move have been in that everyone’s system supports it but not everyone will necessarily be able to use it.

    A fundamental flaw already lay with the fact that the system already excluded people whose internet connections were spotty or restrictive or at worst absent- the way members of the military were effectively being treated has already been ballyhooed quite a bit, so I won’t go into another diatribe with that.

    Even if those facing restrictions could still make it online on a daily basis for the check, how many of them would find their games cut off or diminished for the lack of a connection to the cloud? How many developers then would support such a feature that, possibly, less than half of their active players might be able to utilize?

    I agree that it would have likely seen better use before MS recanted their DRM stance and this change of position will likely hurt the cloud for the foreseeable future more than it will help it.

    The last query I would posit is a question of how effectively all this would run in theory versus in practice? How much bandwidth does this sort of process eat up (an important question if you’re in a connection restrictive environment such as, say, Australia) and how is the game experience affected if your connection suddenly lags or drops?

    Some of these things I ask simply because I actually don’t know.

    I’ve had greater misgivings about Sony’s plans to stream PS3 games in lieu of backwards compatibility, as I know how good my connection in an uptown Los Angeles area using the best (read: only) available internet service provider available to me can be.

    Forgive me if I lack faith at this stage.

  2. The sad thing is it’s very conceptual at this stage. So while I’m about to list some nice examples of shit you would find interesting for using it, I feel like it’s not the most solid conversation.

    For running the world beyond the player’s cone of vision: it’s not really a world if it doesn’t exist until the player interacts with it. It’s a theme park.

    The simplest way this impacts something is in Skyrim. In Skyrim, traders magically restock inventory and gold supplies every in-game day. Their inventory has a chance to be updated with interesting goods based on a variety of things.

    But this economy is very static, since the player is the only one who spends money, who buys things, who sells things. There are no supply lines to disrupt, no markets to saturate, nothing.

    Were the game being fully rendered, you could imagine trade caravans going to and from each of the holds, cities and towns. Bandits or dragons attacking these, while the player is off raising sneak by walking in a corner, would disrupt the economy. This would be an opportunity for Radiant Ai / Radiant Quest design to come into play, as the player is tasked with stabilizing trade routes or carrying shipments himself.

    But you’ve already played a game with a more simplistic “whole world render” model: Majora’s Mask. It was done by creating schedules for all the NPCs, and only needed to be managed over 3 days of in-game time, which was only a few hours of gameplay when scaled.


    Yeah, it would need a near-constant internet connection, though that’s a kind of DRM where it’d be easy to say you’re getting much more than you’re giving. Unless it worked by providing something optional, like the difference between PhysX support for the Batman: Akrham series. No internet connection and newspapers fall in stacks. With it and they flutter around in fully rendered glory.

    But then, it would need to be real and not imaginary, like the recent Sim City fiasco.

    How much bandwidth? Not much. All it’s doing is crunching numbers. On a different scale, our currently existing architecture can handle this for massively multiplayer games, and that deals with collating hundreds of player locations and inputs and redistributing them after performing calculations on them.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: