Gary Marshall, operations director at engineering consultancy Sweco, stresses the importance of innovative techniques that provide additional resilience, while reducing data centre operational costs
In a recent report, research firm IDC predicted that by the end of this decade, the total amount of data globally will be about 40 zettabytes, which to put the number into context is 40 followed by 222 zeros. With this data becoming more and more imperative to business operations, there is greater emphasis on the need for data centres to be as resilient as possible.
With the exponential growth of data in recent years, fuelled in part by what was described by The Economist as the ‘Third Industrial Revolution’, the reliance on data centres for business critical functions has also increased dramatically. In today’s environment, expectations are increasing for mission critical sites to experience no downtime and be fully operational, 24 hours a day, 365 days a year.
Any interruption to power supplies, no matter how small, can cause severe disruption to business operations, including reduced output and – in some sectors, such as financial services – the potential for monetary penalties. More and more businesses are also offsetting their data; just think how much useful work you can do with no access to your network and its data.
While the focus remains on the implementation of highly resilient data centres and reliable storage of information that is imperative to businesses, each company has its own individual set of needs and concerns.
Understanding the needs of a business when it comes to its mission critical sites is vital, with reliability, resilience, flexibility and adaptability all being key concerns. However, one of the most pressing matters in today’s climate is the efficient use of energy, from both an environmental and cost perspective.
These pressing issues are leading companies to seek innovative ways to reduce the environmental impact of their data centres. Many are looking at ways of reducing the energy costs associated with running these critical business assets, while also ensuring they are powered by a reliable energy source.
Rising power demands
Power demands and costs are continuing to rise and it has been predicted that the amount of energy consumed by the world’s data centres will treble in the next decade, putting an enormous strain on energy supplies. According to recent research from the United Nations DESA, Cisco IEA, by 2050 more than 2.5 billion people will live in cities, while the internet of things (IoT) and digitalisation will allow for approximately 50 billion connected devices. Those megatrends, coupled with industrialisation, are driving uplift in energy demand by more than 50% by 2050.
In this scenario, we need natural and sustainable solutions. However, these must be coupled with renewable energy sources and systems that are simplified, flexible and scalable, meeting both current and future expandability in power demands. In the UK in particular, increased energy demand is a very significant issue. Businesses are relying on a single electricity grid, which has seen spare capacity dwindle in the recent past, and which could put the reliability of mission critical data centres at risk.
Sweco is investigating the reduction in power usage energy (PUE) through the use of alternative and innovative solutions for providing free cooling for the majority of the year. This is done by analysing the climate conditions and maximising the temperature bands for electronic data processing equipment.
Energy efficiency can be maximised through the use of permeable façade systems that allow for controlled natural ventilation and through the creation of electronic data processing equipment that operate at much higher temperatures, negating the need for cooling and natural ventilation.
Cooling: innovative approaches
Free cooling has been investigated by other consultants and Sweco is looking at more radical solutions by using the planet’s resources and natural environmental solutions that are all around us. One of the most significant changes in recent years is the way data centres are cooled, and there is rightly a focus on improving these systems. We have seen traditional methods displaced by innovative approaches, with the reduction of energy costs at the forefront of the developments. A particular area of interest is the possibility of free cooling for data centres, and research is under way to improve the use of natural cooling that can be accessed in more temperate climates, such as the UK and mainland Europe.
By designing data centres that can make use of natural cooling, significant reductions can be made in the amount of power these mission critical sites use. This could be so effective that 80% of cooling will effectively be free, with just 20% supplemented by more traditional energy usage. This kind of innovative design can help to deliver highly efficient, reliable and economical thermal management.
A realistic acceptance from clients regarding the input temperature range to the server rooms, which matches the manufacturer’s maximum limits, is essential. Considerations around the life expectancy of the electronic data equipment should also be addressed.
Why not explore electronic data processing equipment that can operate at much higher temperatures? This will negate the need for cooling and natural ventilation is all you need. I guess to an extent, this is gradually happening and we will no doubt see further step leaps in technology over the next five years.
The speed of technological advancements and the rapid growth of the reliance on data mean the design and engineering behind mission critical services need to catch up. Cooling is just one example of the innovative design required in the modern world and illustrates the importance of rewriting 20th century templates and developing new 21st century solutions to suit the global urbanisation trend and the increased importance businesses place on systems critical to their operation.
The trends in data centre cooling point towards evaporative cooling, direct fresh air using ventilated facades and lower PUE using indirect adiabatic cooling. This is a method that reduces server room heat through a change in air pressure caused by volume expansion. We need to look beyond the conventional, explore new ideas and embrace technology. We still only consider resilience in terms of power continuity, but if a design is over complicated it will fail through human error.
Focusing on controls
The next largest contributor to power failure is the controls, so why not double up on these systems? We have developed the SRM 200 module, which doubles the resilience of a transformer/generator control system. Also, circuit breakers are only deigned as either ‘on or off’ – why?
Can we not design a switch that has multiple poles on a rotating shaft? Distribution schematics need to be developed to embrace the new innovations in technology and provide simple, robust cooling and power solutions. Why do engineers complicate the designs by installing complicated PLC systems with hot standby PLCs when they are all backed up with UPS modules, which have exponential curves, so you invariably have hours of battery autonomy?
Other innovative solutions include the possibility of powering a data centre entirely by fuel cells built into the server racks, which has been proposed by various manufactures, and even the ability to generate your own power with chicken litter and burn it to create steam which in turn can drive the steam turbine and create your own on site power station. Other solutions to consider are the use of off peak electricity and battery storage to support both the data centre and/or the national grid for frequency response.
Coupled with emerging advances in technology, the national grid is changing and the spare generating capacity to meet demand has also changed from circa 25% to 6%, so data centres can play a role in meeting this demand and frequency response, from utilising the large generating infrastructure of the data centre and feeding back into the grid.
Engineering companies, including Sweco, are working to redefine current templates, through the identification of risk areas and the development of new solutions. We are currently in the middle of a research programme looking into free cooling using different solutions for varying climatic conditions.
Moreover, we are re-examining cooling solutions to reduce the PUE by going back to nature and learning from the things and resources around us, within the natural world, relooking at the controls for chillers, pumps, AHUs, UPS modules and standby diesel generators. As energy costs continue to increase, we need to consider alternative means of power such as fuel cells, battery storage, photovoltaics, wind energy and below-ground resources.
A further area for consideration beyond individual data centre resilience is ‘cross centre resilience’, which considers the loss of data between various back-up and duplicate sites. If data centres have mirror images, this would mean looking at the whole data centre resilience picture, rather than just considering each data centre individually, but this is a whole different story. These will help to eradicate single points of failure, provide maintainable systems, and minimise the risk of system failures.
Providing cost-effective and reliable solutions using our experience and expertise can ensure the creation of data centre systems that meet the world’s ever-increasing demand for data.