Friday, August 21, 2015

An Insight into ICT’s Energy Consumption and its Implications

[This is a short introduction to the paper of the same title that was presented at "The Fourth Annual Kathmandu Conference on Nepal and Himalaya" on 23 July,2015. The full paper can be found at Martin Chautari's website [Link to Paper].]

The recently floated ICT policy draft and the IT policy of 2010 have both missed to address the energy demand of the proposed infrastructure and service developments beyond hinting it as a barrier. The wholesale developments in the ICT infrastructure to support online services that caters virtually every aspect of the society can bring the power infrastructure to its knees and by its own. There are fundamental limits to how far and fast the efficiency of the digital technology can evolve. The potential of ICT to improve the energy scenario assumes that the digital technology will be situated in every aspect of our lives to extract every tiny drop of efficiency from transportation, power and electricity, telecommunication to lighting our streets. The idea can be described as a move from the material to immaterial resource.

There are obvious flaws to such arguments for a country like Nepal. The saving of critical resources such as the energy due to the adoption of ICT and its efficient use will be buried by the growth that is expected from the IT and ICT policies themselves. This is not a new observed effect in the realm of technology and market dynamics. Since the early days of computing, ICT sector has seen massive improvement in the amount of energy consumed by computations. However, the necessity of the increase in the processing power and the exponential increase in the number of computers over the years have well and truly cancelled the gains over the years. The energy (specifically electricity) required by the computations have overtaken the efficiency in computing. Alarmingly the “Green by IT/ICT" and “Green in IT/ICT" enthusiasts have failed to give due attention to this rebound effect.

The optimism placed in the digital technology to self-improve at a pace in order to be sustainable relies on technological assumptions some of which require the fundamental limits to be pushed by technological innovations. The improvement of the semiconductor industry is closely tied with the economics of the computer industry. The improvement in the technology to meet the doubling effect of number of electronic components inside the chip requires huge research investment. It is widely acknowledged that as long as the industry is getting the returns, the empirical interpretation of the Moore’s law is likely to continue. However, there are numerous challenges to such claims, that the halt in the pace is nearer than expected (as soon as 2036), as they rely on the transistors size to shrink beyond the current fundamental limits.Similarly the demand of the users’ for a faster broadband Internet and communications have been pushing the fundamental limits of the optical fibres specifically the Shannon’s limit Ellis.

These challenges to the trends, the regression effects and the halt in efficiency gains as the technology approaches the fundamental limits is likely to be seen in our lifetime and not in next-generation or beyond. These should have serious implications on the shape and size of the policy interventions regarding situating ICT as the key driver to economic growth expecting it to deliver sustainable growth while these unanswered questions are on the verge of bringing the house down.

No comments:

Post a Comment