3pointD.com interviews IBM
3pointD.com has more on that mainframe project, and I still don’t get it, because of some rather odd statements in the interview.
It will “rely on the Cell’s processor for rendering.” Uh, a metaverse with a processor requirement for clients? Or do they mean server-side rendering, which seems unlikely and a bad idea anyway?
“We see this technology moving into banking and retail and anything where the consumer is involved in a transaction of commerce that they would today do over the Web, online shopping, online banking.” Banking? Does anyone at all want to do their banking in a virtual world? What value-add does a virtual world bring to doing banking? Being able to interact with a menu on an ATM was a vast improvement in efficiency and accuracy for the typical person going to a bank.
Their big concern? “The problem is that rendering is kind of weak. We haven’t figured out how to accelerate that yet, and how to marry that to transactions.” Rendering is the weak point for online worlds? Despite the fact that it’s been shown over and over again that customers just don’t give a damn about the visual representation? On the other hand, there’s a comment about clients for mobile phones, music players, TVs, and so on. Or maybe they mean something like timeliness of rendering given latency, given their comment about transactions…
From stuff within the article, it seems that this is both a hardware solution and a product suite designed to run on that hardware — mention is made of Hoplon’s messaging and physics solutions, of a billing system, and so on. I’m not clear on to what degree you are expected to write your own virtual world atop all this, or whether IBM is planning on creating the server software too.
13 Responses to “3pointD.com interviews IBM”
Sorry, the comment form is closed at this time.
: “This is a brand new way to support the needs of… ¶ SLionhead Blog, Virtual World Branding. on More on IBM’s Mainframe for Virtual Worlds: “This is a brand new way to support the needs of… ¶ Raph’s Website on More on IBM’s Mainframe for Virtual Worlds: 3pointD.com has more on that mainframe project, and I still don’t…
Gartner Says 80 Percent of Internet Users Will Have A “Second Life” by 2011 (Feedster on: second life) 04/26 23:01 IM, Skype, AIM, Yahoo Messenger, ICQ Meet Second Life, Meet Web 2.0 (Feedster on: second life) 04/26 23:00 3pointD.com interviews IBM (Feedster on: metaverse) 04/26 22:59 Why Gaia is the next MySpace (Feedster on: second life) 04/26 22:58 showing my /age (Feedster on: second life) 04/26 22:52 IBM Adds Gaming Chip to Mainframe
Okay, now I agree they’re just playing “buzz word bingo”. On-server “rendering” is mostly pointless, beyond the grossest level for physics (unless they’re really committed to creating an SL-like world that can be viewed on cell phones, in which case smartphone CPU’s will catch up before bandwidth and latency does).
–Dave
IBM’s server middleware software, WebSphere, is the plumbing mentioned in the article and will dictate much of the server’s architecture and implementation language. The rest (the details of the virtual environment) is probably left to the customers to create given IBM’s expectation to work with multiple clients. IMO, they’re poking around seeing if there’s a vertical in the making here for future mainframe sales. Banking connection comes quite naturally given it’s IBM we are talking about.
It sounds to me like it’s possible that there’s something really really cool in IBM’s back room, but the guys answering the questions just don’t have the technical underpinning to truly explain it properly without a lot of buzzwords.
Or not.
Mmm, I wouldnt mind having the ability to pay my bills from within WoW, that would save me approximately 5 minutes per month of tabbing out, starting up a browser and logging on to my banks website.
I also think there might be some interesting business related to selling hardware to people who enjoy destroying money through gambling. There are many such peopl around the world and most of them dont mind destroying money a bit here and a bit there to be able to tell their grandchildren stories about how they were around pioneering various interesting industries.
“When I was young I helped Henry Ford make the car into something everyone could get hold of. It cost a lot of good moniez but look now at all these cars and I am their proud and forgotten creator. Remember grandpa as the man who made the car when you grow up!”
The hardware things are cool. WebSphere is a giant dog. No one in their right mind would write MMO software on top of such a thing. It might be a good choice for large business-to-business solutions or something, but this sounds like clueless marketroids in IBM trying to push it as some kind of MMO middleware.
“Or do they mean server-side rendering…”
It seems that the machine was originally designed for CG and streaming video services, so perhaps they actually do mean server side rendering. Maybe they originally planned to push the hardware for live “TV” shows with synthespians and personal POV streaming video but with the flurry of SL media coverage they decided to change the spin and didn’t quite get it right.
Websphere is okay if you have fully asynchronous transaction model (like web selling, for example). But without major reworking, i can’t see it being effective for synchronous transfers with any kind of latency protection.
What I should have said about WebSphere is that it’s horrifically big and complicated. (Like a lot of “enterprise” products, I suppose).
Yes it can do some nifty things, but you’ll have to hire a full-time WebSphere wizard guy to set it all up. I’m struggling to imagine an MMO design where the benefits of that sort of thing would outweigh the huge drawbacks. MMO servers ought to be low-latency and efficient, which I don’t think anything built on top of WebSphere can really be. =)
>Does anyone at all want to do their banking in a virtual world?
Yes.
There’s a few companies working this actually, I think mostly because of the ability to lower the end-user system requirements and reduce reliance on specific platforms. You’d need a LOT of distributed power to handle very many different users all needing rendering from their specific point of view of course. Maybe these cell processors can do that? I have no idea.
My thoughts on Mainframes For Virtual Worlds, and realism in games.