Scientific Sense Podcast

Monday, October 28, 2019


The future of cloud computing is getting cloudy. The roundtrip from mainframes to personal computers and back to centralized computing has been inefficient, to say the least. It just allowed a plethora of mediocre companies and ideas to try and die. Massive computing power never solved any problems, it just misled a lot of technologists seeking fame and riches. And, in the process, it contributed to worsening the climate problem. Granted, it did create the world's richest person, in anticipation of a never-ending scale-up, that is unlikely to materialize. As Silicon Valley burns from fires started by electric wires, it is time to refocus on computing with less power.

We are reentering a regime governed by distributed computing once again. This time, it is not going to be on desktops but everywhere. It is not going to be about data but decisions. Humans could have taken a clue from their own societies thriving on distributed brainpower. Those seeking efficiencies and scale always preferred centralization (1) not only in computing but also in organizational structures. But with centralization came a variety of costs including but not limited to lack of redundancy, flexibility and, volatile decision-making. Aided by a few monopolistic behemoths willing to sink billions of unused cash on computer farms, the "cloud," has been growing. Their strategies are ably aided by consulting gurus, experts of the present and not the future. Not to be left behind, the developing countries have been in hot pursuit, assembling centralized computing power as if there is no tomorrow.

The future will not require such stranded investments spewing heat and pollution. Instead, we will need to invent massively distributed computing that requires almost no power. The minuscule amount of power needed should be produced in-situ by movement, ambient temperature or air.


No comments:

Post a Comment