Forget the Telecosm, It's the Avanex-cosm
Some of us joke about how much Gilder like Avanex. It seems like most of his reports these days contain about 8-pages on Avanex, and one little aside that discretely promotes a new company to the Telecosm Table. Much to George's chagrin, I imagine, he has learned that he can only add Avanex to the Telecosm once.
But while we may joke about it, Gilder is very serious about Avanex's position in his paradigm. Avanex gets eight pages in every Gilder Telecosm Report because Gilder thinks Avanex deserves eight pages in each Gilder Report. To put it bluntly, Avanex is the Telecosm. Or maybe, just maybe, we're moving past Gilder's old Telecosm and into the new, Avanex-inspired paradigm. As Gilder said at his speech at dinner last night: Welcome to the Lamdbasphere. Gilder and his team believe, and believe fervently, that the success of any optical company going forward depends almost entirerly on its ability to play in the Avanex-inspired Lambdasphere. That's the paradigm.
So What Is the Lambdasphere? And why Avanex?
The Lambdasphere is a world in which we can cram a ton of individual optical wavelengths onto a single fiber. It is a world, as Simon Cao, CTO of Avanex told us in the opening session, in which we use those infinite wavelengths to create a switchless optical network. And Avanex is the company that makes the Lambdasphere possible, by offering the capability to cram hundreds of thousands of lambdas, or wavelengths, onto a single optical fiber. Moreover, Avanex promises the ability to steer those wavelengths through the network using all-optical technologies. To give you an idea of how big this promise is, consider that modern optics works with about 40, or maybe 80, wavelengths on a fiber. And consider that steering those wavelengths requires robust optical-electronic-optical switches. Avanex goes all-optical.
But Why Do We Care? What Can We Do With All Those Wavelengths?
The question, of course, is whether we really need all those wavelenghts. And the answer, of course, is that we will need them if we have them.
Imagine you were designing a network from scratch. What would that netowrk do? Well, it would offer unlimited bandwidth. It would be always on. It would be cheap. And it would be perfectly fast.
Gilder argues, and I think quite correctly, that the only way we get to that perfect network is if we build a "switchless" network. A dumb network. A network where a signal leaves point A and travels on its own to point B, without any electrical reconfiguration and without any expensive switching.
Put another way, the best network is a network that travels between two points. In a point-A-to-point-B network, all the information that you want can travel down its own dedicated conduit. That conduit doesn't have to do anything - just sit there and let the signal flow. Nothing can go wrong. The network is perfectly dumb. It's a freight train on a single track. That's the ideal. As soon as you add a switch to the network you get delays, unreliability, and bandwidth limitations. It loses that freight train simplicity, and starts working like a train in a stock yard. No one can go very fast.
Of course, if you try to multiply that point-to-point netowrk out to a million points, you get chaos. It's just not practical. You can't, for instance, have a million cables coming into one home. Fiber costs money and it takes up space. A million ifbers just won't fit.
The interesting thing about the lamdbasphere, however, is that you don't need a separate cable for each point-of-presence. You don't need a million cables. After all, with the enormous proliferation of lambdas, it's not the fiber that carriers the information ... it's the light. And light is very, very small.
In fact, when you start to stuff hundreds of thousands of lambdas onto a single fiber (this is a number of years off), that physical fiber becomes almost irrelevant to the degree fo connectivity. All of a sudden, you can virtually connect those points-of-presence to one another. The physical limitations are almost moot. All of a sudden we can have infinite connectivity.
Avanex! Avanex! Avanex!
It would be impossible to overstate the importance of Avanex's promise of lambda-multiplying technology to the Gilder paradigm. And if Gilder is right, it would be hard to overstate Avanex's importance to the future of the network. Avanex's technology is quite simply at the core of the way Gilder evaluates optical technologies - How well will it play in the Lambdasphere?. And Gilder's enthusiasm for Avanex is contagious.
If we truly do get all of these lamdbas, we can move towards a state of infinite connectivity - where every point is virtually connected to every other point, without any switches. It's the robust, switchless network. It's total connectivity.
The question, in the mean time, is how Avanex plays in the transition. How do Avanex's technologies fit into the network prior to its reconfiguration as an infinite-Lambdasphere? Gilder and Co assures us, and Avanex's growth rate backs him up, that it fits very well. But for the short-term, at least, that's the question. For the long-term, if you believe Gilder, Avanex is the Telecosm.
---------------
Matt Hougan
MetaMarkets.com
community.metamarkets.com |
