Our Uncertain Streaming Future

One of the more curious things to emerge from E3 this year - and something that remains a curiosity even now that the dust is well-settled - is the seeming consensus that game streaming is inevitably, inexorably making its way towards being the industry's dominant way to distribute games.

Executives aside, not too many people actually seem terribly happy about this notion; but there's a pretty broad acceptance that it's just how things are. Almost every discussion of next-generation consoles is now accompanied by chin-scratching over precisely what role streaming is going to play, and to what extent it'll offset the need for ultra-powerful hardware - a pretty damn radical proposition that seems to be receiving surprisingly little questioning.

Let me put my cards on the table here - I have some skin in this game. Since finally upgrading to a 4K TV, I have of necessity become a consumer of a rather archaic commodity - Blu-Ray discs. I honestly never imagined myself going back to physical media for movies or TV shows, but 4K has been the last straw for my internet connection. It was just about able to eke out a half-decent Netflix stream on a smaller, lower-resolution TV (as long as you didn't mind things going all Lego-blocks now and then), but on the larger screen it smears artifacted pixels around in a manner that gives me flashbacks to RealPlayer's stoic but utterly misguided attempts to stream video over dial-up connections in the 1990s.

"A game you play for 20, 50 or even 100 hours would chew through vastly more bandwidth than just downloading the game to play locally would require"

I live in central Tokyo, not some broadband wasteland. This is the highest-end broadband connection available to me - there is literally no upgrade option I can take short of actually moving to a new apartment with different infrastructure- and it won't stream 4K (and honestly struggles a lot with HD) between the hours of six and midnight every day. My response to people making blasé throwaway comments about the inevitability of streaming, then, is a little more than a raised eyebrow - because I'm damned sure I'm not alone in this. For all the failings of Japan's broadband infrastructure, it's still far beyond that in many other countries - and at least doesn't impose data caps and limits, as is common for many people in the USA and elsewhere.

Now, at this juncture it's worth noting that I'm talking about the broadband situation right now, and the people predicting an inevitable streaming future at talking about, well, the future. They imply that it'll happen within a short space of time, but it's important to define what we mean by a short space of time - because to be totally fair to their arguments, they're not talking about a year or even two years' time. Perhaps people get the wrong impression about how soon streaming will be a major factor from the experiments that platform holders are already running - PlayStation Now is a thing which already exists, and Microsoft is muttering about Xbox games you can play on any device - but I don't think there's any serious, credible suggestion that either of those companies, or any publishers, expects the next Xbox or the next PlayStation to be primarily a streaming device, or even to make streaming into a central part of its offering.

We're talking, then, about the next-next-gen; about having another generation of the kind of powerful devices that play games locally and that we all know and love, and then after that moving to some kind of thin-client setup, or an approximation of that. All-streaming, all the time; which sounds radical when you say it like that, but take a step back and it's hard not to notice that we're really just dealing with the latest iteration of "this is the last console generation, because X is here and it's going to kill consoles", with streaming merely replacing smartphones as this generation's new X.

"PlayStation Now offers access to libraries of older titles, but I have yet to hear a compelling argument for streaming's advantages with regard to newer titles"

That timeframe also makes these predictions pretty tricky, because that's a seven to ten year gap, and a hell of a lot can and will change in seven to ten years. I know it's something of a tech cliché to throw out the fact that smartphones didn't exist ten years ago, but it's true and it's very relevant; smartphones have changed a lot of things but from a purely technological point of view, thinking about where processing is located, they've driven a revolution in high-powered, energy-efficient chips that has put extraordinarily capable computers into most people's pockets, a notion that would have been hugely far-fetched not so long ago. Making confident predictions about a decade-hence streaming future makes a lot of assumptions that could easily be upset by changes in the tech and economics surrounding processing, storage and networks over the coming decade.

Once again giving the benefit of the doubt, however, let's assume a linear development in all of those things. In that case, sure; in seven to ten years there will be a lot less people in my position. Fibre networks will be further upgraded and 5G will be rolled out; the technological ability to deliver high quality, low latency streaming to a significant proportion of the market (still not everyone, and the industry is going to have to think long and hard about how many consumers it's willing to leave on the shelf) will be there. The business models of the carriers remains a wildcard, though. I could, hypothetically, stream 4K video over my LTE connection right now, and the quality would likely be much better than my highly contended fibre-optic connection. LTE connections, however, are capped at a pretty low rate - so even now, for some people the issue is a business one, not a technological one.

That distinction is important, because I get a strong sense that many people are treating this shift towards streaming as being inevitable simply because they assume that it will follow the same pattern as digital distribution - a move which was looming, waiting to happen, and simply needed the technological aspects to fall into place before it became the dominant paradigm. Digital distribution taking over from physical was pretty damned inevitable as soon as the direction of travel of tech was clear; the advantages all across the value chain were simply too big to be ignored.

Streaming is a different proposition; the business case for it remains unconvincing, which essentially means that the technology isn't the real issue. We could quite easily reach a point where streaming is trivial given the tech available, but still isn't a popular way to play most games; I have zero doubt that it will find important applications in some cases (PlayStation Now offering access to libraries of older titles is a great example), but have yet to hear a compelling argument for its advantages with regard to newer titles and existing gaming audiences.

"Right now, everyone wants to be Netflix - but their technological model simply doesn't make sense for a non-linear medium."

That's the thing; you can make a case for streaming in a number of outlier situations, but once you start saying "we're going to move the industry en masse to this model", you need to start thinking in broader economic terms. Here, the toughest, thickest wall this idea runs into is the simple fact that in general over recent decades the costs of processing and storage have dropped far, far faster than network costs - a trend which seems highly likely to continue. This means that storing and running a game in the cloud may be insanely cheap, but beaming that game over the network to players may be very expensive to one party or the other. Anyone who's worked with cloud services to any significant extent knows that that's where the costs really start to add up - network ingress and egress - and consumers may well also face additional costs that make this whole thing unattractive even if it actually works.

Raising this issue with streaming advocates always results in them pointing to the success of other cloud services - which, I think, betrays the extent to which this thinking is informed by a general sense that "music and movies do this now, games must be next". The thing is, this issue of network transport isn't a major hurdle for existing cloud services, because they generally fall into two camps - small files (documents, photos, music) which don't impact bandwidth much, and can be cached locally easily anyway; and large files (TV shows, movies) which take up a bunch of bandwidth, but are generally only watched once, which means the bandwidth cost to stream is no more, and likely somewhat less, than it would be to download.

Games are a radically different proposition from a network perspective. Never mind the question of latency - let's assume that's solved in ten years' time. However, by the nature of how games work, a game you play for 20, 50 or even 100 hours would chew through vastly more bandwidth than just downloading the game to play locally would require. Sure, you can play it on a device that's low powered because all it needs to do is display the stream; but here again the question is, why?

Processing power is cheap, not just in the cloud but for consumers too; the same goes for storage. Both processing and storage continue to pack more power into smaller devices at a pace of progress that far outstrips the build-out of network infrastructure. Streaming seeks to turn the weakest link in the tech chain, the network, into the lynchpin of the whole experience; it puts the industry and its consumers at the mercy of broadband infrastructure companies, betting the future on the fervent hope that their costs will fall and performance rise at a dramatically greater rate than has been the case in previous decades.

From an industry point of view, I do understand the temptation of streaming - no more piracy, total control, games as a service - but unless these things make sense for business and consumers alike, the idea isn't going to take off. The benefits cited for consumers aren't entirely inconsequential; games that start without needing a download, no need to buy an expensive console device, access to a huge library, and so on - but they don't stack up to an offering compelling enough to outbalance the looming question of just whose pocket all of these bandwidth costs will come out of.

Right now, everyone wants to be Netflix - but their technological model simply doesn't make sense for a non-linear medium.