Otoy says supercomputer will enable revolutionary games-on-demand service

If you’re wondering which games-on-demand company is the real one, today’s news isn’t going to help you out.

While Steve Perlman’s OnLive has revealed concrete plans to launch its service today, another rival, Otoy, is revealing today that its video compression technology will be the foundation of a new kind of supercomputer that will enable the digital distribution of video games.

Jules Urbach, chief executive of Otoy (right), is making the announcement today at GamesBeat@GDC, our executive game conference in San Francisco. His company clearly has a ways to go, but he is making up for that by partnering with powerful companies.

The multiple-teraflop supercomputer will be made by Supermicro, a leading maker of high-end servers. Together with its support from chip maker Advanced Micro Devices, the relationship with Supermicro is the strongest endorsement yet of what has been an experimental technology at Otoy. And in another interesting twist, which Otoy is staying mum about, another one of its partners is Intel.

Supermicro says it will launch Fusion Render Cloud Servers in the second quarter. The Fusion Render Cloud is a new kind of supercomputer that uses a bunch of servers with both microprocessors and graphics chips in them. These servers can be used to create high-end graphics imagery that can be piped at lightning speed to a home computer. The effect is that a low-end computer could be used to display high-end games, thanks to the processing that happens in the so-called cloud.

Otoy’s ORBX codec — the compression software —  is one of the key links here because it allows one server to send thousands of streams of high-definition video or game graphics to just about any display. The Los Angeles startup was created by graphics technology wizard Urbach, whose topic today is Disruptive Innovation at GamesBeat@GDC. In an interview, Urbach said consumers will be able to use the technology this summer.

One of the cool features of Otoy is progressive downloading. This means you can operate a digital distribution business such as Valve’s Steam network without having to wait a long time for downloading. As with OnLive, you can step into the play immediately. You may be playing a game on a server at first, but it eventually downloads to your computer where you can play it locally. Of course, if your machine is a low-end machine, you don’t have to play it locally. Urbach says the company has tested this technology with 400 games.

As with its rival OnLive, this cloud-based computing breaks loose a digital distribution business model for games, which are typically sold for $60 in retail stores. With this technology, web-based companies will be able to sell games directly to consumers.

About 10 supercomputers could probably support a million users, Urbach said. For this reason, AMD views the Fusion Render Cloud as the next evolutionary step in cloud computing.

“We look at Otoy as a technology that we can use to reorganize everything around cloud services,” said Charles Boswell, director of digital media and entertainment at AMD.

With these announcements, Urbach’s dream of getting a whole alliance of companies to launch a games-on-demand service is coming together. He thinks that services using Otoy will be launched for consumers starting in the summer. You’ll be able to play fast-action games such as Left 4 Dead 2 on any device by tapping into the Fusion Render Cloud, which is enabled by the supercomputer.

The consumer service will be similar to what Steve Perlman envisions with OnLive, the well-financed games-on-demand service backed by big game publishers and AT&T. But Otoy’s approach is very different. The company is operating in a horizontal fashion, while OnLive is more vertical, doing each layer of the service itself. Otoy is licensing its technology to be used in the supercomputer, which is fueled by graphics and processors from Advanced Micro Devices. Hosting companies will offer the cloud-based service to publishers of games and other apps. And consumers will ultimately subscribe to the services.

The hardware itself is quite intimidating. A supercomputer will consist of 128 servers, with a total of 250 AMD “Mangy Cours” Opteron microprocessors and 500 graphics chips based on AMD’s Cypress designs. Each of those graphics chips can process 2.7 teraflops, or 2.7 trillion math operations per second. Each supercomputer could serve 3,000 high-definition users, or 12,000 standard-definition users. Otoy’s own software on a consumer’s own machine is tiny, taking up just four kilobytes of data.

The system will work with Mac, PC, Linux, Android, Windows Mobile, iPhone,  iPad and custom set top box solutions based on cheap ARM chips. Urbach says that Otoy is trying to strike bundling deals now that will allow its software to reside on a wide range of platforms.

This horizontal approach, where each company provides just a layer of the overall solution, could allow Otoy to reach a larger group of consumers and companies than it otherwise might, Urbach said. In that sense, the games-on-demand service enabled by Otoy will resemble something like Amazon’s EC2 service, which leases cloud computing and storage to startups.

The Otoy technology isn’t a pure phantom. Dassault Systems has shown its SolidWorks application, a 3D modeling and design program, working on a netbook via Otoy. Normally, that software has to run on a beefy workstation, but with Otoy it can run on any device. Urbach has also shown high-end games running on iPhones using the technology.

Compared to OnLive, Urbach is certainly the underdog here. He has been talking about delivering this disruptive technology since 2004. But people took notice when Otoy announced the AMD alliance in January, 2009. OnLive has positioned itself as closer to a consumer launch than Otoy, although it’s possible that both may be out in the market at the same time.

The key question is whether Otoy, OnLive or a third rival, Gaikai, can pull off server-based gaming for millions of consumers on the same network. If they can, they will turn the video game industry upside down by cutting out retailers and eliminating the need to buy high-end game hardware.

As we’ve said before, the goal of shifting computing from the client (i.e., the user’s machine) to the server is an inexorable trend. Google hopes to undermine Microsoft’s client-based Office software with server-based Google Docs software. If it succeeds, Google could disrupt Microsoft’s 500-million user base.

Doing something similar in the games market — creating video games that can be instantly played from a server with no download to a client machine — could be similarly disruptive. If you can buy a game over a network, you don’t need the retailer. And if the heavy computing is done on the server, you don’t need anything more than a display in the user’s device. There is no need for a gamer PC or a game console. You can also play your games on any machine as long as you log into the same account via the Internet. The cost of delivering these games will likely be small, which means consumers could enjoy more games for less money.

These benefits are so obvious, you would think companies would have done it already. But technically, it’s an enormously difficult problem. Faster broadband connections help, delivering better throughput, or one-way traffic speed. But games need low latency, meaning fast interactions back and forth. Improving broadband speed helps with throughput but not with latency.

Urbach was a co-founder of The Groove Alliance, a Flash game maker that made a splash in the 1990s with cool 3D web games. He also started LightStage, a facial capture company that helped capture human faces so they could be easily animated in movies such as Brad Pitt’s The Curious Case of Benjamin Button. LightStage, coincidentally, does work that is similar to that done by another Perlman company, Mova. Both were used in the Benjamin Button movie, but for slightly different effects.

One of Otoy’s investors is George Gilder, the one-time pundit who advocated the onset of the “Telecosm,” a world awash in bandwidth, in the days before the last bubble. Gilder has been one of the true believers in Urbach’s vision, in part because Gilder himself publicized the whole notion — articulated first by Eric Schmidt, who was then at Sun Microsystems and now is the CEO of Google — that huge broadband connectivity would “hollow out the PC,” meaning that a world with so much networking connectivity would have no need for heavy-duty user computers.