P A P E R T I T L E
The Future Potential
A U T H O R / C R E D I T S
A B S T R A C T
Two inexorable metatrends are driving the telecommunications industry, and will result in significant increases in the size of, and demands on, the III - V class of semiconductors. Both key trends result from digitization of everything. The first metatrend is a series of convergences: voice with data at the device and network levels, wireless with wireless networks, and network architectures converging on packet switching. The second metatrend is an insatiable demand for bandwidth, motivated by the technical demands of digitized multimedia, by the human need for speed, and by cost of capacity.
Convergence and bandwidth have direct implications for networks and terminal devices. There will be more coax and fiber on the wireline side. To keep pace on the wireless side, we see the rapid rise of large markets at high frequencies - anywhere from 2.4 gHz to 60 GHz. There will also be much more embedded wireless mobility, caused by the blurring of device categories and wireless-wireline convergence. The Bluetooth Group, for example, is sorting out some of the basic architectural issues. Finally, the inevitability of lower costs and prices from the action of Moore's law of digital things means quantities will continue to grow rapidly.
The fundamentals of very significant demand increases are coming into place. This presents some interesting challenges; at least three deserve special note. First, there will need to be technical rationalization of the various flavors and design approaches to III - V semiconductors; they won't be discrete or special purpose parts. They could become mainstream integrated devices. Second, we would like to see much more integration, including resistor, inductor, and capacitor elements. The higher the frequency, the more long circuit traces present design and packaging problems. This will represent major technical challenges. Finally, we think GaAs will see challenges. One of these is the new class of silicon germanium HBT's which may offer some interesting competition at "lower" high frequency. Another is the indium phosphides class which may offer potential at the "higher" frequencies.