Too Much Throughput?

2:00 PM -- The history of high tech is just riddled with statements about great leaps forward in technology not being necessary. 10 Mbit/s? Whatever would one do with that? And let's not forget Bill Gates's infamous observation that 640KB should be enough memory for anyone.

So it's amusing to hear some analysts talk about .11n as having performance well beyond the requirements of any application. Again, there is no .11n yet, but 100 Mbit/s is likely to be the lower bound on realizable throughput. Farpoint Group's testing of today's MIMO products shows that some of these (residential-class, mind you) products can indeed achieve that speed. We've got parity with wired Ethernet! But does anyone need it, he asked, knowingly.

Like Ethernet, wireless LANs use a shared medium. While I would argue that raw throughput alone is enough of a justification to upgrade, sharing implies the need to get one's traffic on and off the medium as soon as possible, thereby freeing up capacity for those we share the airwaves with. And, at least for asynchronous data, one can never really have performance that's too fast.

But the real reason for all this additional capacity is to support time bounded traffic, most notably voice-over-WiFi, but also video, assuring enough headroom for decent service. So, MIMO (and eventually .11n) make the best use of the scare resource that is the radio spectrum; they improve throughput; they allow for much better time-bounded traffic than is otherwise possible; and they get more users on the air per unit of frequency and time. Too much throughput? Never!

— Craig Mathias is Principal Analyst at the Farpoint Group , an advisory firm specializing in wireless communications and mobile computing. Special to Unstrung

Be the first to post a comment regarding this story.
Sign In