Very interesting. Check this out:
seas.upenn.edu
As far as the switching speed, I believe that it is fine. The important thing isn't the speed of the switch but more the idea that you can do the switching without the need of electronics. The full bandwidth and lowest latency is ensured by keeping everything optical the entire way.
Think of it like this: there's a big universe of light with gargantuan amounts of bandwidth already out there that is being hooked up to slow (electronic) switching equipment..
I've been waiting since 1994 to find this idea of just wasting bandwidth in someone else's brain besides my own. I guess I should have done more searching..
This basically takes all the brains out of terminals, and allows them (the brains) to be centralized.
So for example, one possibility would be to standardize computer I/O devices (monitors, keyboard, etc) and interface them via fiber to computers that are hundreds or even thousands of miles away. It would be possible for example to get a kind of "direct access" to any computer on the network rather than effectively having two memory spaces (one for the machine running locally [terminal], and one for the machine running remotely [server]) as it is done now.
For another analogy, consider something like "Java." In order to run Java code, your own computer has to do all this processing. And you have to make sure java is compatible with your OS and hardware etc etc. But if instead of doing all that crap, you just get a fiber with all the output that the human eye can handle (there is a limit -- see any psyc book) and if it can run on the server, it works. This means that people won't have to "maintain" their computers anymore, won't have to install hard drives. Won't have to install applications, etc. It's really a big thing. And if you make something new that works on any computer compatible with the output that every terminal has, then anybody can use it regardless. In this way, the abstraction of computer hardware will evolve more rapidly since we can break free of all computer architecture issues (so long as it involves a monitor, keyboard, and whatever other I/O devices, the brains could be anything [EG: roving camera with remote control and computer augmented knowledge]) but it still "looks" like a computer to anybody who uses it.
This would allow you to run any operating system by effectively having a big digital computer switchbox that would allow you to connect to any computer, and not just throw a few bytes around, but have it send you full-bandwidth video data.
It's one of those ideas that really powerful. So far, I've explained this concept to two people: my wife (as she videotaped me because she thought I was so off my rocker and couldn't understand what I was talking about [when I first thought about the possibilities]) and a co-worker of mine (who kept sticking with the adding a bunch of electronics -- didn't understand that you just want to solve things with bandwidth instead of transistor smarts.
Back when I first realized all this (in 1994) I basically freaked out and broke out into a "song" called "break the bandwidth" for the video camera. Now I'm going to have to break that thing out and watch it!
Ok .. I know I'm weird, but this really excites me.
I'll probably wake up and realize this is all just a dream!
Standard disclaimers apply -- I have a Bachelors in Computer Science and a minor in Electrical Engineering from Texas A&M .. graduated with honors .. have all the background, but I could just be hallucinating (though I think this is what I've been waiting for as far as getting in early on the next big thing -- assuming 99% of other people have never though about all of this stuff before yesterday and _everybody_ will be thinking about it very soon)...
We'll soon see news stories about how this is shaking up the whole economy just like we saw stories on Integrated Circuits back in the late 70's and early 80's. But I do think that this is even bigger (if you can believe it).
So.. I'm long on "A"
--
- Aaron |
