Cloud Gaming 2012: The future is overcast
As different players try to define cloud gaming, where does the technology go from here?
The Cloud Gaming conference kicked off today, with participants discussing the state of cloud gaming and where it's headed. Cloud gaming, it seems, is confusing. Presenters had several different definitions of what the term cloud gaming means, and there was disagreement over the status and the future of cloud gaming and its various components. There was general agreement on only one point: Cloud gaming is important, and it's becoming increasingly important.
One useful way to measure the importance of cloud gaming to key companies in the industry is to look at the number of employees they have attending. Companies may talk about priorities, but where they spend their money illustrates their priorities more clearly. Looking at the employee counts, there are 3 from Nintendo, 3 from Amazon, 4 from Activision, 6 from EA, 8 from Ubisoft, 8 from Sony/Gaikai, and 11 from Microsoft. This might indicate the level of interest in cloud gaming, and which companies might be implementing it in the future.
The conference started with a keynote by Michael Pachter, analyst with Wedbush Securities, who generated some controversy when he maintained that the free-to-play business model is not sustainable in the long term. He feels that publishers should be able to get paid in some way for every player of their games; this might be from advertising, sponsorship, subscription or payment. Others in the audience disagreed, saying that they and others have generated healthy revenues from free-to-play games. Asia was cited as a region where free-to-play was especially popular, and without it piracy would make revenue unlikely; besides, putting in ads would drive players away. Pachter felt that such concerns about advertising are ill-founded; "If players pay something you can provide ad-free gaming, and if they don't pay why should you care if they stick around?"
"If players pay something you can provide ad-free gaming, and if they don't pay why should you care if they stick around?"Michael Pachter
Next, Pachter sat down for an on-stage chat with Frank Gibeau, president of EA Labels. Gibeau refused to reveal anything about the existence of next-gen consoles, despite Pachter's insistent questions. Questioned by Pachter about the reasons for EA's poor stock price, Gibeau said "The whole sector is under pressure right now. There's a real lack of clarity about what's happening in games." As far as mobile goes, though, EA is bullish. "We're the number one publisher on smartphones and tablets, probably the best-kept secret in the industry. We intend on doubling our share there and growing very aggressively." As far as free-to-play goes, the vast majority of EA's revenue from mobile titles comes from free-to-play. Gibeau noted that "Last quarter almost 20 per cent of our revenue came from free-to-play."
Turning to Zynga, Pachter asked jokingly if EA was going to hire back all of those ex-EA employees that recently left Zynga. More seriously, Pachter wondered what lessons could be learned from Zynga's decline, and what needs to be done differently in social games. Gibeau pulled no punches: "We were always focused on being a multiplatform company, platform agnostic is the saying we have around EA. Being solely focused on one platform is a high-risk business; you can have things happen like just happened to them, in terms of an overhyped opportunity, and a little bit of headwinds happen and the next thing you know you fall really fast and you fall really far. The cautionary tale is something we always knew, which was the best bet is to be a multi-platform company and to have a very broad base of intellectual property."
Pachter then queried Gibeau about Star Wars: The Old Republic, which was not as successful as EA had expected, and their subsequent rapid shift to free-to-play. "I think it was a really good experience for us to go through," said Gibeau. "We feel like the business model is out of position for where the market is right now, based on subscriber churn." That wasn't all he had to say on the subject. "People were a bit price-sensitive to the subscription," Gibeau admitted. "We knew at some point in the future we were going to move Star Wars to the free-to-play model; we basically accelerated those plans and brought that forward. In fact, the audience reaction in our subscriber base has been very positive."
"Being solely focused on one platform is a high-risk business"Frank Gibeau
Pachter concluded by saying that "I think the next gen consoles are coming out, and I think the console business will be fine. I think that it's the changing business models, the emerging business models for social and mobile, that are going to define the industry for the rest of the decade."
Next, Peter Relan, executive chairman of Agawi (formerly iSwifter) took the stage to explain his theory that cloud gaming is nearing an inflection point, where it may increase rapidly in acceptance as the necessary market conditions are met. Relan defines those as a true GPU Cloud (such as Nvidia's GeForce Grid), a universal player (similar to how Netflix plays on a variety of plaforms) and good controllers (ones that resemble console controllers with joysticks and buttons). Relan feels that all of these elements are happening now or in the near future, and with the right business models cloud gaming is poised to grow rapidly once these elements are in place.
Ron Haberman, CEO of startup CiiNOW, has a different view of the proper way to implement cloud gaming. He has built a service that aims to be invisible to users; his customers will be publishers, network operators, and other large companies. He sees the best opportunity as being similar to Amazon Web Services, which powers many popular web sites behind the scenes. CiiNOW plans to provide a seeamless back end, operating on a wide variety of hardware, for companies that want to provide streaming games.
When Valve's director of business development, Jason Holtman, took the stage he defined cloud gaming as something different than the other speakers. Holtman sees cloud gaming as a new set of connections and relationships between players, communities, and developers. "Some of the best ideas will emerge from frameworks relying on cloud," noted Holtman. "Fundamentally different games are possible and will be expected. Doing this relies on open, connected and malleable systems."
"Paradoxically pixel streaming will only work as long as it doesn't become popular"Osman Kent
The next technical session discussed a very different vision of cloud gaming. The co-founder and CEO of GPU pioneer 3DLabs, Osman Kent, spoke about his new companies Numecent (a cloud service for software) and its spinoff Approxy (for cloud gaming). Kent spent time explaining why the 'traditional' cloud gaming model is fundamentally broken. What Kent calls pixel streaming (the common conception of cloud gaming, such as Onlive employs) has fundamental problems. It simply cannot scale in a cost-effective way, Kent said. Problems include not just bandwidth requirement, but the fact that network connections are stable and reliable enough for the demands of gaming. Issues like jitter, contention, throttling, data caps, ping distance, and latency are all difficult to manage, and prevent many users from being able to handle pixel streaming. Huge capital expenses makes the business model dysfunctional, according to Kent, as the OnLive implosion clearly demonstrates. "Paradoxically pixel streaming will only work as long as it doesn't become popular," Kent said.
Kent's solution (which of course Approxy is built around) is something he calls cloud paging - the push-pull paging of instructions, not pixels, using local GPU resources (in your PC) to handle the processing. Essentially, Approxy sends just the code necessary to execute what needs to happen at any given point in the game, which Kent claims can be done with only 5 to 10 percent of the total footprint of the game. Kent says this approach is 1000x more efficient and scalable than pixel streaming.
Technical issues aside, Kent closed with a warning about the state of cloud gaming: "We ignore the IP issues at our peril; within the next 12 months this will become a patent minefield," he said. If indeed cloud gaming becomes hugely popular, thus generating large revenues for somebody, litigation is likely to follow.