Wednesday, June 9, 2010

Wireless data fails through its own success

On Monday, Steve Jobs had his iPhone 4 demo fail due to Wi-Fi congestion problems at the WWDC conference. (Blogger Liz Gannes of GigaOM theorizes that ad hoc hotspots set up by attendees contributed to the problem.)

Also this week, while passing through O’Hare airport, I desperately tried to log onto the Boingo paid hotspot but could never get a reliable connection among the dozens (hundreds?) of travelers also trying to use Wi-Fi.

There’s been a lot of publicity about AT&T cutting back on unlimited all-you-can eat bandwidth for its 3G service — thanks to heavy usage by a significant fraction of the iPhone customers — as well as the speculation that other 3G carriers will do likewise. The problem isn’t going to go away with LTE (or WiMax).

But this is also clearly a problem with Wi-Fi as well. Networks designed for people checking email or flipping 2 web pages a minute cannot handle the traffic generated by YouTube.

What the wireless industry seems to ignore — probably deliberately — is that this is an inherent (and perhaps insolvable) problem of wireless communications. Wired communications are scale free — you can add new wires and routers and switches indefinitely.

On the other hand, wireless spectrum cannot be extended indefinitely, and there’s only so much smaller you can make the cell sizes (from microcells to picocells). The past decade has benefitted from increasing spectral efficiency, but spectral efficiency in 4G is only slightly better than 3G.

So while Moore’s Law seems like it will have a 50 year run, the improvement in wireless data efficiency appears it will only run about 20 years before it hits the wall.

Has the industry overpromised? Is it ready for the consequences when it fails to get the additional spectrum it dreams of? Right now I see the problem, but not the complete implications or possible solutions.

1 comment:

Mike Demler said...

I have seen the wireless industry put some emphasis on the inherent limits of wireless communications, making that the central issue in lobbying the FCC for more spectrum. At CTIA Wireless, AT&T's Ralph de la Vega made that point and the FCC's Genachowski agreed. (To the chagrin of the broadcasters, when he suggested re-allocating unused TV bandwidth).

Bandwidth is not infinitely expandable of course, but there are other potential remedies. One of the most interesting is Cognitive Radio, which can enable dynamic spectrum reallocation. This would require a lot of changes in the industry and infrastructure, but some interesting design work is already underway.