I am constantly amazed by those folks who always seem to question the latest advances in network performance, most commonly throughput. This goes way back. Did you say 10 Mbits/second? Who could possibly need that? (Ethernet, circa 1975.) The same was true for 100-Mbit Ethernet and Gig-E. Of course, these arguments are now viewed as being as silly as Bill Gates' original pronouncement that 640 kbytes ought to be more than enough memory for mere mortals. At least we in the wireless world have the good sense to charge prices sufficiently high that this excess throughput is kept well out of reach of those who might need it. Pay $80 a month for EV-DO or UMTS? I don't think so.
But the bottom line, I think, is that we'll see the same thing on wireless as we did on wire-to wit, too much is never enough. Given the scarce resource that spectrum is becoming-even the 2.4-GHz band is getting congested in many places-we should as a matter of policy seek to minimize the amount of time any given transmission takes. That's where higher data throughput becomes essential. It's a bit like buying a faster processor for your PC-you end up optimizing for idle time. But more idle time on the air lets everyone get more done.
Let's look beyond the mundane applications of today and move forward a couple of years to the upcoming era of wireless media. Have you looked at the High Definition Multimedia Interface (HDMI) that's now appearing on many home entertainment devices? It provides up to 5 Gbits/s, which is enough to go all the way to 1,080p resolution. Just imagine how good the constant parade of TV commercials is going to look! On the other hand, think about connecting a projector to a DVD player on the other side of the room. Priced any HDMI (or other) cables lately? And running said cables-well, that's one of the reasons I got into wireless.
One hope for provisioning this amount of bandwidth is in using the 60-GHz unlicensed band. Yes, 60 GHz is tough to design and fab; we're talking gallium arsenide today, but I've recently had some encouraging conversations about using silicon germanium and even CMOS here. Yes, this is the oxygen-absorption band, but we're not talking really long range. Yes, millimeter waves are very directional, but innovations in antennas march on. Will 60 GHz become a fixture in future multimedia links? I'd count on it.
But this is not to say that UWB or even MIMO-based WLANs won't be able to compete here. New JPEG2000-based compression parts (check out ADI's ADV202, for example) could provide really, really good video at high resolution over links with significantly less throughput. And just the savings in time from not needing to install and configure all the wire required in even a modest home theater will increase consumer satisfaction, lower support costs and make the decor-conscious everywhere positively swoon.
And this brings me back to my original point. If we've got the wireless bandwidth for Viagra ads in high-def, we've got the wireless bandwidth for almost any data requirement on the horizon. I remember just a few years ago (OK, 12) waiting for e-mail to download on a 4.8-kbit/s Mobidem. But as we've learned from the history of microprocessors, it's a bad idea to underestimate what engineers can do, and even worse to argue for limitations in the technology.
Craig Mathias (craig@farpointgroup.com), principal at Farpoint Group (Ashland, Mass.)