Skip to content Skip to footer

Tips to combat rising costs & temperatures in the data centre

Chris Wellfair

Chris Wellfair

Projects Director at Secure I.T. Environments
Share on facebook
Share on twitter
Share on linkedin
sdecoret
Image credit: sdecoret / Shutterstock.com

Data centre managers are in unprecedented times now, and things on the horizon look pretty gloomy. 

Last year’s COP26 conference outlined in the starkest terms what lay ahead if nations did not start to make serious progress on breaking their drug dependency on fossil fuels and countless other areas that are negatively impacting our environment and future.  This year in the UK alone, we are being reminded of those words, with the highest ever recorded temperatures of 40.3 degrees celsius, a whole 1.8 degrees higher than the previous record in 2019.  That is a huge increase.

If that were not enough to contend with in the data centre, keeping the data centre running, and cool, is now more expensive than ever. Wholesale market energy prices rose from £1,042 in summer 2020, to around £1,971 earlier in the summer, and are expected to top £3,000 at the start of 2023. In fact, at the start of August, energy consultants Cornwall Insight said that a typical annual domestic gas and electricity bill in England, Wales and Scotland could reach as high as £3,615 for an average household in the new year.

The problems for the data centre itself are very real when it comes to temperature control, as we saw with Google and Oracle cloud services, and countless other UK data centres struggling with the impact of the heat wave.

So, how do you square the challenge of keeping down energy consumption and costs while cooling an environment that is getting hotter? Investing in new more efficient cooling is one way to go, but that is not going to help you today or tomorrow. Here are some top tips from on areas to check, that might highlight an opportunity to reduce energy consumption and improve the efficiency of your cooling.

  • Review your EMS rules – Over time, we make tweaks to the rules on the EMS in order to meet the changing needs of the DC. A wholesale review of all the rules may identify opportunities for efficiency where rules are either redundant, or perhaps create conflicts.
  • Use your DC data – Intelligent PDUs, environmental sensors and utilisation data, can tell you a great deal about what is happening in the data centre, but only if you look at it! We see lots of examples of data centres where this data is used to make purchasing decisions, or identify immediate or impending faults, but the reality is it can tell you so much more about where efficiency gains can be made.  If you don’t have a dashboard that brings all this data together, get one, but more importantly make the time to analyse it and look for opportunities.
  • Maintenance – Is all your maintenance being conducted correctly. This includes moving rubbish that could impact air flow or block vents. Are vent filters cleaned regularly and the DC vacuumed? These things might seem trivial, but they matter and change the efficiency of cooling in the data centre. Also check that the maintenance regime for the cooling systems themselves is up-to-date for the current physical infrastructure and being followed.
  • Lighting control – What types of lighting do you use in the DC? If it is not LED, then this is one area that could make a big difference to consumption. If you don’t have the resources to support this (though it is relatively inexpensive), look at other opportunities. Could you add PIR sensors to detect when lighting should be on in different aisles, or change the lighting schedules?
  • Server consolidation and optimisation – Are there servers running in the DC that are running minimal loads, and could these be consolidated in some way to reduce the energy consumption and cooling requirements, without impacting resilience? Are there servers that are only required to provide services at certain times? These should be shut down or hibernating outside those windows?
  • Change the DC cooling temperature – Depending on your individual circumstances there may be an opportunity to change the environmental temperature in the DC. Even half a degree could make a big difference to energy consumption, particularly at cooler times of year. This kind of change is best undertaken once you know the data you are getting from your EMS is accurate and allows you predict what scope there is and how it might impact individual servers and cabinets, particularly when under load.

Plan for the longer term

You’d be surprised how much of a difference the tips above can make to your energy consumption and cooling: they are quick to implement, relatively low cost, and can make an immediate impact when time and resources may be a little stretched. But they can only get you so far, and if on balance you are looking at your HVAC infrastructure and know you have been putting off the inevitable for too long, then maybe it is time to get a plan together. It may be that over time, the way your infrastructure has grown means that you need to reconsider the current arrangement of your hot and cold air flow in the data centre.

The efficiency of air conditioning units is improving all the time, and over the last five years alone there have been some dramatic technology advances that lower energy consumption, while improving cooling, and work more intelligently as part of your overall environmental management system. Free cooling, for example, brings outside air in to flow through the data centre, rather than re-cooling the same air source in a closed environment. Liquid cooling too, can be excellent, providing direct to processor cooling capabilities, and is especially well suited to high performance computing environments.

Changing your whole cooling infrastructure is of course not a quick or inexpensive project, and for data centre managers, the challenge is often making the case for such large capital investments. Ultimately, all any company has to do is look at what has happened to the big players with the big budgets, to see it can happen to anyone. For data centre managers, hitting a brick wall, the board just needs to think about how long they can survive and the cost of having no, or limited, IT services for a few days. 

Feeling the heat

Striking the balance on efficient cooling, whilst managing costs is a tough one, especially when capital resources for investment are limited. But these tips should help you make more of what you have. Whilst it is very difficult to say when the energy crisis will come to an end, it is reaching a pivotal point at the end of this year as nations decrease their reliance on gas and oil from Russia. Every data centre manager needs to be looking now at every opportunity to reduce their energy consumption with the resources and creative thinking they have available.

Show CommentsClose Comments

Leave a comment