Green Indexes Used in CAST to Measure Energy Consumption of Code

This blog is from CAST’s keynote speech at MeGSuS’16, 3rd International Workshop on Measurement and Metrics for Green and Sustainable Software. Download the presentation here.
Fueled by our growing thirst for constant connectivity and the dawn of the Internet of Things, the energy required to power all the world’s computers, data storage and communications networks is expected to double by 2020 according to the latest research by McKinsey & Company. This would increase the total impact of IT technology, in terms of global carbon emissions, by at least 3%.

Making The Case For Energy Efficient Code

The current state of measuring the environmental impact of our IT infrastructure is missing a big piece of the puzzle. One of the metrics we use, power usage effectiveness (PUE), only looks at how much power entering a data center is being consumed by the computer hardware in relation to the total amount of energy the facility uses.
But what about the millions of lines of code running on that hardware? How can we know if that’s energy efficient code?

CAST Tries To Save the Planet with Green IT Index

It’s simple physics: a piece of application code gets caught in a logic loop, the CPU heats up as the increased throughput tries to make sense of the commands, the computer reacts by pumping more power to the motherboard and cooling system to keep everything up and running, and your electricity bill goes up.
For a household server, this increase might be a miniscule blip in your monthly bill. But for a large IT organization with hundreds of thousands of servers located in datacenters around the world, that could start eating into quarterly profits.