Presentation from Peter Levine at A16z on the theme of edge computing will see a return to distributed computing the in the years ahead.
- The amount of data collected/created at the network edge is vast
- You must process locally and then upload summaries.
- Device complexity and capability is increasing rapidly e.g. Cars, Smartphones. This supports an intelligent edge.
- Local processing supports faster decisions, better feedback loops.
The end result is that ‘Cloud Computing’, whether private or public becomes a centre to ML/Data Analysis.
Takes about 25 mins to watch.
The EtherealMind View
FWIW, I have had this view for sometime and I agree with line of thinking. The public/private cloud is good for some activities and many others will be distributed.
My example is a highly-automated factory with large numbers of robots. It is impractical to backhaul data from thousands of sensors per robot to a centralised systems for decision making. It must be processed locally for reliable decision making.