How is it possible that a power company based in the technology mecca of the world is unable to turn power off where it is absolutely needed?
Edge Computing Can Keep the Lights On…
…and power critical medical devices, care centers’ operations, and refrigerators — not to mention California’s many electric vehicles for which the state gives generous rebates
By James Fox
Nearly one million California residents are without power as I write this thanks to PG&E, which in response to anticipated wind conditions (that never really materialized), has brought down the grid across much of Northern California, from the Bay Area to the foothills. The shutdown started Oct. 9 across 22 counties. According to a PG&E vice president, the measure is a last resort effort to prevent wildfires.
But his answer merely prompts more questions.
How is this a solution? What about upgrading the infrastructure? How is it possible that a power company based in the technology mecca of the world and the most progressive state in the country is unable to turn power off discretely to only those areas where real danger is present? It’s a quandary, particularly when the answer lies in a technology solution that exists today.
This doesn’t need to be the new normal. If your head is in the Cloud, you may not be aware of the next evolution in data storage and collection: Edge computing, where critical data is collected at the source through smart devices. Imagine this scenario: Mega volumes of data are collected real-time at ultra-fast speeds, filtered and analyzed, and acted on via a time-series database imbedded in sensors installed on utility poles to monitor wind conditions.
No imagination necessary: Time series databases are able to monitor power and transmission lines locally and constantly target problems on the grid in sections measured by meters, not hundreds of miles. In this way, the grid can be powered down strategically and predictably, impacting only specific areas versus dozens of counties.
Edge computing is a relatively new concept in Silicon Valley, but in Asia it is a different story. There, MachBase has successfully monitored hundreds of thousands to millions of data points a second on devices as small as a Raspberry Pi, enabling real- time analytics at the point where the data is generated.
As this massive blackout — which customers in Asia have figured out — demonstrates, we don’t have time to send data back to the cloud for big data analysis and decision-making. We need edge computing “boots on the ground” and local predicting and targeting. Edge computing is changing how we monitor smart factory robots in real-time, are able to collect 100,000 records a second per robot, and is increasingly essential in steel mills, ships on the ocean, CCTV cameras, and yes, data lakes for energy grids. By 2020 (next year!), International Data Corp predicts that 45% of all data created by IoT devices will be stored, processed, analyzed, and acted upon at the edge.
PG&E: Rethink fire prevention. The answer is already out there. #Machbase
James Fox is the vice president of strategic partnerships for Machbase.com, a time-series database company for edge managed platforms, with offices in the Bay Area.