Ken,
"Frank, I would appreciate your comments and opinion on this article..."
The first half of the article was too loose, and had some technical flaws. To me, I found it downright silly in some parts. I could hardly find the incentive to finish it, but I did. But only for you, Ken. grin
Even some half-truths - that would ordinarily pass the test of plausibility for works of prose - are botched, as the very license that plausibility allows is misapplied. Unless I'm asked to, I wont even bother to list the grievances that I registered mentally during the read.
The one element that I did find noteworthy and *not* off base, is the implied direction that technology is pointed in. To be clear on what I'm referring to, towards an all optical (or mostly optical) network. Then again it's not too difficult among like-minded folks to point a finger skyward and suggest: "I think that's where Heaven is at."
But directions alone don't necessarily get you to the end state, nor do they suggest that the end state will be reached by Labor Day of this year, or next, or the next. Especially in such areas as replacing end user packet flows with direct beams of light.
Every time I've been engaged in a discussion on this topic I've asked for an engineering reference model, or a reference design, that demonstrates how it would scale beyond a handful of users at this point in time, or even in the near future. I never, ever get a response. At one point I had to go out and find one myself, and I did. It's name is SONATA [go to google: "SONATA switchless"], a European optical network initiative. But it does not portend to do the type of things that the author points to, specifically.
Don't get me wrong. I'm not stating that end-to-end all-optical will never work. In fact, if I were a betting man I'd put money down stating that it will eventually work at some point in time. Core first, then the edge, and then out to the end point (which is the final stage before end to end can work). All I'm saying, however, is this: If you can tell me how it will work and what the rules of administration and addressing etc. will be, and then show me a product that will support all of these wonderful features, then just show me! If you can't, then come back in a couple of years when you can. Simple.
And while you're waiting for the "end-to-end" o-o-o model to become soup, don't stand there knocking and ridiculing, ad nauseam, the only infrastructure we've got that's capable of work-horsing our data loads around.
I'm often amused by how some journalists - while doing their must-carry stories...
[you know the type... mesh will win over ring; SONET is evil; ATM is dead; the PSTN sucks; now IP sucks, too; copper is worth less than spit unless it happens to be copper in coaxial cable that's being enhanced with a wireless protocol; IP routing is an anachronism, since bandwidth will be free, but caching <why is it needed if bandwidth is free?> with our cache is good; yes, bandwidth will be free, but here's a magnificent new streaming video compression scheme; etc.]
...can both celebrate a new technological direction, and at the same time find it necessary to traduce the very technologies that necessarily came before it in the natural order of evolution.
Imagine someone stepping on your shoulders to hop over a fence without saying thank you, but instead leaves you standing there while calling you names from the other side? C'mon.
What's even more surprising to me is when a prospective superior technolgy attempts to make an inferior technology better, in the name breakthrough. Applying certain wireless protocols to black coaxial TV plant comes to mind, when the obvious superior alternative is to go with fiber, which is a triple irony if one believes in -spheres, -igms and -cosms.
To net it out, from my flying altitude and air velocity [which are intentionally low and slow, respectively, to catch more of the detail without the effects of image-motion blur taking place (Hi Grace)], I found it difficult to keep a serious perspective during some parts of the article, to answer your question. Yes, all very large plays will eventually take their lumps. Equilibrium almost demands it. The re-ascent of IBM comes to mind here, as personified by several stories that featured Big Blue during this past week. Not too shabby, eh? But look where they were when they were taking their lumps.
The most dubious notion of all is to suggest that the top-tiered IP plays would not mobilize to compete (it's by no means too late, since many of the techs that threaten aren't even into their chasm stages, yet), even if through acquiring, in order to fend off the legions of upstarts who possess the technologies that are being referred to. I suppose the only way to put an end to such noise would be for a very large IP company to acquire the optical play in question, or one of the others in a growing number of plays who are in the same league.
And there are now, and will continue to be additional others, in the same or more advanced leagues, as times goes by. That's the way this stuff has always worked.
FAC |
