Have we been living in a fool's paradise with regard to the Internet? Do global power considerations mean that the Internet is about to become a limited resource and much more expensive, even taxed?
It looks likely that global chip sales will decline in 2015 compared with 2014 implying that average selling price erosion is exceeding unit supply increases. A decline in the value of the global chip market in the absence of an economic crash, a natural catastrophe, or at least an oversupply bubble, has been very rare.
One of the problems would seem to be that the smartphone, as killer product and market driver, has more or less run its course. Many people are hoping the next killer application will be the Internet of Things; nanoelectronics everywhere.
Think twice about that, because a nanoelectronics research compendium makes the point that — without several orders of magnitude reduction in power consumption in electronics — the projected roll outs of mobile data and the IoT could cause a global energy crisis as soon as 2020. That's just over four years away. The book is CHIPS 2020 Vol. 2: New Vistas in Nanoelectronics , edited by Bernd Hoefflinger (see Internet blackouts by 2020, warns nanoelectronics book). Mainly a benchmarking of the state-of-the-art in nanoelectronics, the book also makes some important societal points.
It's not just the billions of end-point “leaf nodes” that will be consuming power — probably battery power unless nanoelectronics can make the nodes autonomous — but the exabytes of data being generated and sent to and from data centers. According to Hoefflinger's book that data processing is rising at a compound annual growth rate of 61 percent at present, and is simply not sustainable.
But if it is not sustainable, what is going to give way?
Unfortunately, one of the conclusions in the book is that the traditional evolutionary progress of electronics rarely allows radical solutions to come to market — even though radical solutions are now required to reduce power consumption.
So if the electronics industry cannot prevent the global electronics power budget from increasing rapidly then EITHER many more power stations must be built – but of what type – OR the public's insatiable demand for the Internet must be stifled, probably by higher charges or taxation.
No One Likes Taxes
Building fossil fuel, nuclear and renewable power stations are all fraught with problems and Hoefflinger's conclusion is that the world is unlikely, in the short term, to bring online the quantity of power stations required to meet the forecast Internet demand. So it seems that the future of the Internet is power-constrained.
But taxation and the Internet is traditionally a thorny question. In the early days pledges were made not to tax the Internet for fear of suffocating a new medium that could be a booster of productivity and trade. It has proved to be much more than just a booster, but rather a transformer of global business. And in the process governments have found out that the Internet creates super-national corporations that play one country off against another in attempts to minimize the taxes they pay.
It would be hard, if not impractical, to create an equitable global taxation system for the Internet. But if we do nothing significant to reduce power consumption per node and across the network and allow data volumes to double every 18 months, by 2020 the Internet's energy demand will rival the world's total generation capacity, according to Hoefflinger's book. At that point communities and governments will be forced to introduce rationing and to choose between power for essential infrastructure such as hospitals and transport, power for the lights, and power for the Internet.
Brownouts of the Internet will probably precede blackouts. They will likely manifest themselves as a gradual degradation of quality of service as ISPs are forced to make their own resource allocation decisions. But it can also be imagined that at times of global events such as natural disasters or wars large-scale failures will happen.
At the same time businesses that are based on the Internet, such as Amazon, Facebook, Google, and just about everyone else will be saying this cannot be allowed to happen. There will be thoughts about first- and second-class access to the Internet and about top-down management of peak demand, just as there is with tariffs on electricity.
A Fool's Paradise
Have we been living in a fool's paradise that is about to come to an end with a sickening crunch?
Of course, any such crunch will be mediated by money in some way because as resources become scarce they go up in price. One possible outcome would be significantly higher and progressive charges from Internet Service Providers used to meet government levies to pay for investment in power generation, which is effectively a tax.
The progressive nature of the charges means that small-scale users of the Internet, emailers and so on, will attract low charges but excessive users will be charged at a higher rate.
And lest you think that this is all unlikely, be aware that Pakistan introduced a 14 percent tax on Internet usage in July 2015,; that legislative battles over taxes on the digital economy are being fought in the US Congress and in 2014 public protests forced the scrapping of a draft law in Hungary to levy a fee on each gigabyte of Internet data transferred.
All that said electronics engineers have a clear opportunity and a responsibility to develop energy-efficient solutions for the benefit of future generations. And reading the two books in the CHIPS 2020 series is a good place to start that task.
Article originally posted on EE Times Europe.