Saturday, January 6, 2018

Was Negroponte Wrong?

Lots of things can change in four decades. The Negroponte Switch, for example, suggested that legacy networks made inefficient use of bandwidth. Broadband signals (at that time television) were moved through the air using wireless, while narrowband signals (at that point mostly voice) were moved using cables.

There was another angle, namely that mobile and personal endpoints (phones, at that time) were perversely limited to static fixed network connections, while devices that functioned in a “place-based” manner (television sets) were connected using wireless.

Prof. Negroponte argued we should do just the opposite, namely move narrowband and mobile signals over the air, and confine broadband to cables.

These days, the switch is really from cabled to wireless and mobile, since most traffic now is broadband, and increasingly that traffic is mobile and personal. By perhaps 2019, as much as two thirds of all data will use some form of (mobile, Wi-Fi or other untethered networks, short range or long range), Cisco has predicted.


Of course, assumptions matter. In the 1980s, it would have been impossible to foresee the huge explosion of mobile usage; the shift of TV consumption from place-based to mobile (and from linear to on-demand); the decline of fixed network voice; or the rise of the internet itself.

Nor would it have been possible to accurately foresee the impact of orders of magnitude decreases in the cost of computation and communications. Rather than a shift, we now see a move of virtually all communications to untethered modes.

These days, Wi-Fi is the preferred local connection technology in the office, home and indoor venues. Outdoors and on the go, mobile connections are the norm.

In the new developing areas, such as internet of things apps and sensors, untethered access also is expected to be the norm, not fixed access.

Negroponte was correct--within the limits of networks and costs at the time--in suggesting a shift of broadband to cables and narrowband to wireless.  

Some 40 years later, everything--all media types--are moving to untethered access. That is the result of mobility emerging as the dominant end user device, the growth of Wi-Fi as the universal endpoint connection method, the impact of Moore’s Law on computing and communications costs, the growth of the internet and ever-new ways to use communications spectrum more efficiently.

In the case of millimeter wave and spectrum aggregation, cheap computing means we can use bandwidth assets that were impractical in the past.

Computing power that would have cost $100 million in 1980 cost about $100 in 2010, less than that in 2017. In other words, costs have dropped at least eight orders of magnitude.

Predictions about the future always are perilous. What we have seen is partly a switch, but more profoundly an upheaval. Increasingly, the untethered and mobile networks are for actual access. The fixed network is--in the realm of consumer services--a backhaul network.

No comments:

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...