Energy Efficiency | February 26, 2020

A Fist Full of Information

The Human Brain: A Comparison

The human brain stores a lot of data. The brain has the potential to store as much as a petabyte of information. A computer with that capacity would weigh 293 lbs which is still over 4 cubic feet of space. We do it in the space of a few fists. With the human brain as the standard of efficiency in this case, data centers are going to need a few more decades to shrink to fit.

Data Defined

  1. facts and statistics collected for reference or analysis.
  2. the quantities, characters, or symbols on which operations are performed by a computer, being stored and transmitted in the form of electrical signals and recorded on magnetic, optical, or mechanical recording media.
  3. things known or assumed as facts, making the basis of reasoning or calculation

Definition from lexico.com

Data is everywhere. Every click, every conversation, the photos, the apps, the predictive text as you type…. Everywhere. And it’s stored somewhere.

General Facts About Data Centers

There are over 8,500,000 data centers in the world. More than three million of them are here in the US. That’s one data center per every 100 people in the US, and one for every 800,000 people worldwide. Even more impressive, the amount of power used for these data centers equals 2%-3% of the world's energy use. Let that sink in.

Efficiency is a key term as we investigate data centers and their abilities and requirements. Most centers look to upgrade IT equipment every three to five years (Moore's Law says “the number of transistors on a microchip doubles approximately every two years”). Technology is always improving, and infrastructure can revamp congruently. Individual server racks efficiency improvements can have less impact compared to overall efficiency upgrades to temperature and cooling controls in these buildings.

Raising the baseline temperature inside the data center – known as a set point – can save money by reducing the amount of energy used for air conditioning. It's been estimated that data center managers can save 4% in energy costs for every degree of upward change in the set point. On average, these massive buildings spend half of their consumed energy on what is considered ancillary equipment such as lights, cooling systems, humidifiers, and the monitors.

Uptime and Downtime

Uptime matters for data centers. They are storing, gathering and processing information for more than one company in many instances. Multi-tenant data centers have racks, servers or even rooms where individual companies have their information housed. Then there are enterprise data centers, which house the information for a single corporation. Any chain department store that emails you likely has one or two enterprise data center locations.

Another reason that data centers use 416 million terawatts of energy annually is because data centers are active 24/7. They don’t turn off, and if something goes wrong resulting in one being off for any amount of time, even 12 minutes, the cost is egregious.  Ranging from $7,900 to $9,000 (on the lower end) per minute the cost of downtime is crazy, in addition, average downtime situations range from 84 to 134 minutes… Millions of dollars are lost very quickly.

Global and Individual Statistics

It has been estimated that all US data centers improving their energy use by 20% will save roughly $2 billion dollars a year in energy costs. That is approximately 15,163,000,000 kWh which is equivalent to 1,072,085 metric tons of carbon dioxide and the same amount of greenhouse gas emissions as over 1 billion pounds of coal burned.

The numbers exist around data center emissions because they impact everything from our financial transactions to our choice of morning news.

PUE Means Power Usage Effectiveness

Power Usage Effectiveness is a term specific to the scale which demonstrates how efficiently or inefficiently data centers use energy. Data centers power IT equipment, and if they only power the IT equipment the score would be 1. So, when the percent of energy being used to power anything other than IT, the number goes up. An extremely efficient facility would run at about 1.2 PUE. If the number was all the way up to 2.5 then for each watt going to IT equipment there is another 1.5 watts going to ancillary equipment and costs. An average score is around 2 but the research says that if all data centers were running with the most efficient equipment, the average could be closer to 1.6. Many of the legacy centers that we look at are in the 2.5-3.5 range which easily have room for improvement.

A data center in Atlanta worked with us to improve their energy use and are now saving 4.6 million kWh per year. The greenhouse gas emissions saved from 4.6 million kWh is the equivalent of what would be generated from 3,583,681 pounds of coal burned.  This data center started at a PUE of 2.1 and we got them down to 1.5!