90% of data produced and stored is not simply useless; it gets in the way of useful data. Waste data makes it harder and harder to find and make use of good data.
John Booth gave me a great example of how just because you can collect loads and loads of data, it doesn’t mean that you should. “We work with a couple of organizations and some of them were quite proud to tell me that they were collecting 1,440 data points a day on the temperature of a certain item in a data center. And I think, why are you recording 1,440 data points? The temperature variations are so minute. Why not use an average and you can shrink it down to one data point? Or only look at the data points where it has gone out of a system boundary. Only keep what you actually need in order to determine why that temperature rose, what was actually happening for that temperature to rise. I think they kind of agree with me, but sometimes it’s just easier to keep it all.”
There has never been more data and never less understanding of how to manage it properly. “We as a people have kind of forgotten what data is,” John says. “And we’re kind of hoarding data for no real reason. You have to have an annual spring clean. The energy costs of data are going to rise. Those conversations will need to be had at senior level, and a decision will have to be made to look at the data, see what’s useful, see what’s not useful and take steps to get rid of it. Or convert it onto tapes or CDs. And that is another Green IT concept of Hot, Warm, Cold and Frozen data. So, Hot data is data you access within a month. Warm data is data you may have looked at over six months. Cold data is data you haven’t looked at for a year. And Frozen data is data you haven’t looked at in two years. And of course that CD or tape that will store the Frozen data doesn’t use any energy.”
It can be hundreds if not thousands of times more efficient (less energy intense) to save data in cold storage than in hot storage. The age of cheap data is over. Energy costs are going to keep rising. “A data center in an enterprise can be using up to 40% of the total energy of that enterprise,” John explains. “The financial directors are going: ‘What is my data center doing? You are using 40% of my energy and I don’t know what metrics you are using to justify this.’” How can you justify spending so much money on storing so much crap data? It doesn’t make sense. It happens because of laziness. Because people don’t want to make the effort to clean up their crap data.
Data centers are full of entire systems that were replaced by other systems that were replaced by other systems. These systems were never decommissioned when the new systems were installed because it was easier to leave them sitting there like zombies—the living dead of IT architecture. “We estimated that it was £14,000 to keep one server up and running annually, including all the software and licensing aspects of it. And it’s a zombie server. It’s just sitting there doing nothing. Up to 40% of servers in the US could be zombie servers.” It’s estimated that there are about 70 million servers active worldwide at any one time. (Each one of these servers caused between one and two tons of CO2 in their manufacture.) That means that every year, €1.1 trillion is being spent simply on the servers that store data. With better management—cutting the crap—€1 trillion could be saved every year.
Season’s greetings and health and happiness to everyone in 2023. The next issue of World Wide Waste will be on January 9th 2023.
Podcast with John Booth: Data centers: Data theatre and the tsunami of frivolous data
Interviews with prominent thinkers outlining what can be done to make digital as sustainable as possible.
Listen to episodes