06 May 2021
The pandemic has driven awareness of the critical role data centres have in society. We are so reliant on online services they are as important as public utilities. Robust, reliable and high-speed connectivity everywhere, not just the office, is now the expectation as working and schooling from home continue. If the tip of the internet of things triangle is your mobile phone, the base is the data centre.
For every online shopping click or MS Teams school lesson, there is an energy and environmental impact. The data centre industry has a responsibility, especially when it comes to environmental impact to ensure we are being as efficient as we can be while maintaining utility levels of service.
Data centre operators and managers need a renewed focus on optimisation given the scrutiny the industry is now under. The European Green Deal has led the data centre industry to set up the Climate Neutral Data Centre Pact to ensure data centres are climate neutral by 2030 and supporting the EU goal of Europe being climate neutral by 2050.
These long-term goals are of vital importance, but short-term cost reduction is just one of the positive benefits of focusing on optimisation. Energy costs are at least 25% of the TCO of IT servers and storage solution. These energy costs are expected to double over the next 10 years.
Minimising the energy footprint of IT infrastructure should be a strategic priority for any data centre manager. Professional help is key to achieving the highest standards of optimisation so here are my top 5 tips to get you started.
1. Monitor and measure: know your baseline
Use existing tools. Building Management System (BMS), Data Centre Infrastructure Management (DCIM) or even more simply – the electricity bill – to understand what the data centre’s energy consumption is. The more detailed the better. If you can, calculate the breakdown of the uninterruptible power supply (UPS), cooling system and IT load. It’s key to identify where any weaknesses may lie, as you can compare them to other data centres of that size.
However, if that’s not possible, then you should at the very least understand the total kW and kWh of the site. This baseline is important for comparison and evaluation.
2. Trial and evaluate: make small changes to monitor and evaluate your optimisation
You can do this in three ways:
● Turn up: Data centres across the globe have been gradually raising temperatures as servers and other data centre equipment have become increasingly more robust. Increase the set point of your cooling system by 0.5C° or 1C°. Monitor for seven days and evaluate against your baseline.
● Slow down: decrease the fan speeds of cooling units by 5%. Again, monitor for seven days and evaluate against your baseline.
● Use simple techniques: this can be with tools as simple as your eyes and a piece of paper to find out where air is escaping from your raised floor, for example. Then, simply block the holes!
3. Increase server utilisation:
Raising server utilisation is a major part of enabling data centres to operate at maximum efficiency. Most physical servers only run at about 10% to 15% utilisation according to Pulsant data. This is hugely inefficient, because an idle server can still consume up to 40% of the power of a fully utilised server. Virtualisation is an important technology for boosting server utilisation; in virtual environments, more services run on fewer servers.
4. Identify and decommission comatose servers:
Data centre energy consumption could be significantly reduced by identifying and turning off unused servers. The Uptime Institute has studied this issue for several years and says that “between 10% and 30% of servers are dead and could be turned off. Just by taking full advantage of power management features and turning off unused servers can cut data centre energy requirements by about 20%”. However, identifying servers that aren't being utilised is not as simple as measuring CPU and memory usage. An energy efficiency audit from a trusted partner can help you put a program in place to take care of comatose servers and improve overall server utilisation.
5. Make it airtight: don’t let your money go down the drain
Air flow management can improve data centre efficiency by preventing the recirculation of hot air from IT equipment. Deploying airtight racks and increasing the data centre’s set points help make little changes for big savings:
● Check if the top, side and bottom of the rack have vents, and then block them off. Side cable management causes a lot of bypass air, so use a brush strip or foam to block it. Similarly, the very bottom of a rack can leak a lot of air. Use some foam as preventative maintenance.
● Don’t be over sensitive: find the most thermally sensitive item of IT equipment and understand its thermal operating range. The likelihood is, it will be much higher than you think. A server from 2000 can operate up to 35C°, and you can use this as the basis for the lowest cold air delivery temperature.
As the role of the data centre in society becomes more critical and visible, optimisation and reducing environmental impact must be top of mind. However, by working with trusted partners, there are quick wins that can be made in improving sustainability, managing costs and improving reliability. Whether you need to reduce energy consumption or lower your operational costs, learn more about the performance optimisation services Vertiv offers.
Simon Brady, services channel business development manager, Vertiv
