What Gilder suggested was that routers will be relegated to small footprint CE type devices over smart optical networks. This of course assumes that optical networks are capable of routing via IP since most applications are written with IP networking in mind. So I suppose you can say it's the death of electronic routers but certainly not routers. Traffic will always need to be routed and to suggest otherwise is simply stupid.
Gary, I think that Gilder was suggesting something a bit different. He was predicting an era where advances in DWDM, all-optical switching, and long-haul transmission would allow individual internet users to get dedicated wavelengths on long-haul backbones for the duration of a connection. Thus, on such a backbone, while a routing device would be needed to set up the initial connection, the connection would act as a circuit from then on. So there'd be no need for high-speed routers anywhere except at the edge of these backbones...if you're laughing by now, I don't blame you.
With the costs of optical components being as high as they are, the thought of creating a network at some point within the next 10-15 years that would allow a 56K user to take up an entire OC-48 wavelength (or even an OC-3, for that matter), for two hours so that he/she could download Beethoven's 9th at a blazing 3 kb/s comes across as ludicrous to me. Could it happen someday? Sure, why not? But to expect such a scenario to manifest itself before 2010, if not later, is nothing more than wishful thinking, and thus the possibility of such networks being created shouldn't be factored into the considerations of investors when they analyze the investment potential of companies such as JDS Uniphase, Avanex, Applied Micro Circuits, Cisco, and Juniper.
I hope that no one takes this the wrong way. In some regards, I do have a lot of respect for Gilder's work. However, like Gary, the more that I read his reports, the more I get the feeling that the understanding of intermediate-term (five years or less) technological realities, not to mention long-term business dynamics and consumer tendencies, don't factor into his writings as much as they should. I'm not just talking about his predictions on the triumph of off-the-shelf network processors and wavelength-switched optical networks, but also on his claims regarding the lack of need for intelligent storage, and the future success of cdma2000 vs. W-CDMA, Terayon's modems vs. Broadcom's, and Mirror Image's network vs. Akamai's, not to mention a couple of other topics that I can't think of right now.
Once again, I do agree with a number of Gilder's calls, and I am a shareholder in JDS Uniphase, Broadcom, Applied Micro, Nortel, Qualcomm, and Avanex; it's just that I think it's best for those who read his work to think about the implications of his predictions from an intermdiate-term/long-term financial perspective rather than just a long, long-term technological one before jumping on the bandwagon.
Eric |
