We're all familiar with the relationship between user bandwidth demand and the capabilities of optical-networking equipment and components. Fortunately for the industry, the recent increase in bandwidth-intensive applications such as the Internet has created a requirement for fiber-optic technology in general and higher-capacity optical-networking systems in particular. Equipment for terabit-speed networks represents the most recent product of this relationship. However, a look at current research into tomorrow's fiber-optic technology reveals that forces other than user demand also can fuel the capacity fire. In fact, sometimes that fire appears self-perpetuating.
In other words, one big optical system begets another. This trend is obvious in some cases, such as Nortel Network's recent announcement of its 1.6-Tbit/sec amplification system. If you have a multiplexing system that can transmit 160 channels of 10-Gbit/sec traffic, you're going to need an amplifier with equivalent capability. One can imagine that Nortel also just raised the bar for optical switch and crossconnect manufacturers as well.
The effect the capabilities of one system can have on the research into another isn't always this easy to trace, however. Take the work now underway into 40-Gbit/sec Synchronous Optical Network/Synchronous Digital Hierarchy (SONET/SDH) transmission gear. On the face of it, 40-Gbit/sec transmission appears to be the next logical step in capacity evolution, proceeding from the other SONET/SDH benchmarks that have established what one could consider a natural order of progression. Yet the advent of dense wavelength-division multiplexing (DWDM) calls this order into question. After all, a mere four channels of 10-Gbit/sec traffic squeezed through a multiplexer will produce 40 Gbits/sec of capacity--and if you can run 160 channels at that rate through a DWDM device, you'd have to wonder why you'd need to look at 40-Gbit/sec systems at all within the next five years.
This reasoning would suggest that research into 40-Gbit/sec transmission will prove little more than an academic exercise until one of two things happens. Either system developers will come up with a device that multiplexes 40 channels of 40-Gbit/sec traffic or they'll create a box that transmits a single stream of 40 Gbits/sec more cheaply than a 4-channel 10-Gbits/sec multiplexer can. But anticipated developments in another system area--Internet protocol and Asynchronous Transfer Mode switch routers--may create a need for 40-Gbit/sec devices sooner than the current capabilities of DWDM equipment might lead you to expect. As a representative for a company that has already announced an upcoming 40-Gbit/sec product put it, "What we're hearing from customers is the desire for 40 Gbits in anticipation of the enormous growth in data-capacity demand, and in particular the view that high-speed router technology will be continuing to move to even higher-speed interfaces. So we see that one of the drivers for the 40-Gbit applications is really to provide connectivity between routers as router speeds go up and router interfaces go up to the OC-768-type range."
Thus, we see that ever-higher speeds in one system area can drive equally high speeds somewhere else, even against what would appear to be a strong tide in another direction. Adam Smith's invisible hand is a pushy appendage indeed.
Stephen M. Hardy
Editor in Chief