Intelligent data versus big data

Abstract

The advancement of nanoelectronic functionality and memory capacity has enabled a data explosion, which is promoted as "Big Data", and doubling the mobile Internet traffic every 18 months. In this fashion, the mobile Internet would need over 500 GW of electric power for its operation in 2020 and another 500 GW of embodied power for its manufacturing, maintenance, recycling, and disposal, together one third of the expected total global electric power generation. In the face of this unlikely scenario, a review proposes that this "Big Data" is a result of the artificial, numerical, linear world, inflated by the pocket calculator in the 60s and its successors and now out of control due to linear digital video originated by the CCD imagers since the 70s and mapped onto CMOS imagers. By contrast, we (and our brains) sense our world with a logarithmic response, where minimum response steps record signal increments that are constant ratios of the magnitude of the individual signal. This has been recognized in digital audio, where it led to MP3 as an intelligent data compression. This realism is still missing in digital video, which causes >70 % of the mobile Internet traffic. Transistor-enabled logarithmic recording, invented 1992, and processing of images can compress video data by an order-of-magnitude vs. present mainstream digital video, a significant saving on video traffic and storage. A similar progress from "data" to information can be envisioned towards real data on the "Internet for Everything" like the environment, health, weather, mobility, and robotics.