New IBM Mainframe Platform For Virtual Worlds

The International Herald Tribune breaks the news that IBM is launching a new mainframe platform specifically designed for next-generation virtual worlds and 3D virtual environments. In concert with Brazilian game developer Hoplon, IBM will use the PlayStation3‘s ultra-high-powered Cell processor to create a mainframe architecture that will provide the security, scalability and speed that are currently lacking in 3D environments — a lack that is one of the factors keeping them from becoming widely adopted. If it works, it sounds like worldmakers working on IBM’s platform should be able to support concurrencies far above todays’ capabilities, and implement commerce systems far more secure than is currently possible.

The IHT story talks about a server system that will permit higher levels of concurrency at greater levels of rendering and realism. The machines will be priced beginning at hundreds of thousands dollars, according to the story.

While it probably won’t have much impact on the state of virtual worlds right off, IBM’s new infrastructure could make a big difference in the long run, by enabling much greater numbers of concurrent users in next-gen virtual worlds, and by creating more secure possibilities for commerce. Big media and entertainment companies continue to be interested in virtual worlds, but they are also skeptical in many cases because there is no way to support many thousands of audience members at a single event.

The new mainframe infrastructure also doesn’t answer questions of fragmentation and convergence. Even at higher concurrencies, IBM’s new system doesn’t move the virtual world toward a more unified state, in which only a single browser or client is necessary to access any 3D space. Some other force will have to come along to push that.

Since IBM started poking around in virtual worlds in the middle of last year, it has had its devs, hackers and other geeks quietly peeking under the hood at companies from Linden Lab (makers of Second Life) to Multiverse and more. While the company has been doing a lot of work in Second Life and even some in Multiverse, it hasn’t been very clear just what tack it was planning to take — though its enthusiasm has been plain: for some time now, its execs have been comparing the advent of virtual worlds to the advent of the Web in terms of its impact on business and communications.

The new mainframe architecture looks like IBM’s big bet to capitalize on the shift toward the 3D Internet. Look it for to have a big impact on the state of affairs over the next couple of years.

[Read more on IBM’s new mainframe architecture on 3pointD, including an interview with IBM’s David Gelardi.]


  1. Ordinal Malaprop

    This sort of investment further cements, for me, the idea that IBM really are interested in the potential of VWs beyond the usual corporate PR puffery and me-too-ism – something they have done well in demonstrating in the past. You’re quite right that in itself hardware won’t solve some of the basic problems, but it does indicate that there is significant will behind the project. A company does not launch a new mainframe platform if it isn’t serious.

    What I would be interested in seeing now would be efforts to start discussion on and create common protocols; the punch of a company such as IBM could bring these issues to a prominence beyond the noodling of bloggers and actually get real projects moving.

  2. Prokofy Neva

    >IBM will use the PlayStation3’s ultra-high-powered Cell processor to create a mainframe architecture

    could you explain then how that relates to a thing like the SL download working on a PC or Mac?

  3. Mark Wallace

    @Prok: It increases the processing power of the server, which, if you have a good graphics card, is where lag is created. So a VW built on this system should be able to support hugely greater numbers of concurrent users at higher levels of graphical detail, without lag. Such a VW would still run on a PC or Mac as normal, I imagine.

  4. Trevor F. Smith

    Prok: With that much server horsepower, you could render everything on the server and just send video streams of the world to thin clients, like a personal Destroy TV for everyone without a usable 3D card.

  5. Earnest Candour

    “The machines will be priced beginning at hundreds of thousands dollars, according to the story.”

    Won’t the cost be a little prohibitive for end users? Though I guess it will come down in time. Does this mean that buying an island on one of these servers will cost a fortune, or will they be able to host multiple islands on one server?

  6. Aleister Kronos

    Well we’ve waiting a while to hear this. Yet another step toward 3D internet, and open virtual community. As you say, Mark, without a unified client it is limited, but this is where others (no doubt including IBM themselves) are beavering away.

    The important thing for me is that is takes the servers into a realm where they can stand a chance of managing huge numbers of concurrent users, and deliver them the richness/depth of (user generated!) content that VWs like SL do not have the architecture to support. Also, building basic security and governance features into the framework provides an infrastructure for the effective construction of (something approaching) true virtual citizenship.

  7. Trackback: eightbar
  8. Pingback: SLionhead Blog, Virtual World Branding. » IBM helpt Second Life.
  9. Pingback: UgoTrade » Blog Archive » Developing Worlds, Virtual and Real:Things happen faster than you think!
  10. Pingback: UgoTrade » 2007 » April » 27
  11. Chihiro Hayashi

    Good for IBM, selling their expensive mainframe, and better for Hoplon: it comes with hoplon’s middleware. Bad for the end user, hoplon’s middleware seems like a very low-performance java based app, there are much more efficient and solid networking middlewares for mmo’s around – not to badmouth them, it’s just that technical people in brazil are very sceptical of any technology that may come from hoplon’s game, “taikodom”, which has all the ingredients of a major commercial failure (delayed and confusing development, over-budget, unpopular sci-fi gameplay, and too ambitious for company’s 1st project.)

  12. Pingback: Bob’s Blog » Blog Archive » Sun Aims New 3D Environment at Business
  13. DieStirbMorra

    Taikodom is a vaporware…Tarq and his associates (they own Hoplon) know nothing about games, nothing about person management…Taikodom sux hard!!! Taikodom is vaporware

  14. Shinzon

    I am one of the group of American beta testers recently introduced to the Taikodom world. The game is quite stunning they are just having to deal with network issues I hope IBM can address being that the mainframe is located in Brazil.