As plans proliferate to take data centres “off-world”, are we looking at all the options closer to home first?
Managing power requirements and energy consumption is a significant challenge for the data centre managers of today. With the constantly growing amount of data being stored, created and shared by businesses and consumers alike, the challenge of maximising data centre performance levels whilst minimising expenditure is constantly evolving.
This has arguably been one of the key drivers in an ongoing industry-wide experiment that seems to be growing in popularity – building data centres in extreme environments in order to deliver optimal performance. We have seen the construction of data centres in arctic climates and even the prospect of heading under the sea, but what if the next step is shooting for the stars?
While the decreasing cost of space travel is beginning make this seem possible in the future, there are serious questions to be answered around whether doing so would produce the business benefits to justify such a huge risk.
Whether planning a data centre on the moon or closer to home, engineering simulation has become an invaluable tool.
As external environments become more extreme, the ability to comprehensively simulate a data centre’s performance under all conditions becomes even more important. Simulation enables management to build a picture of how the data centre will respond in a range of scenarios, and tailor performance to reflect specific business goals.
One Giant Leap for Data Centrekind?
The realms of possibility in terms of space occupation are shifting, and this could soon stretch beyond the current array of broadcasting and navigation technologies operating outside our atmosphere.
While the concept of a data centre orbiting the earth is slowly becoming more of a realistic possibility, the first thing we must examine is whether the risk and expenditure involved in developing a space-bound data centre could provide a viable return on investment. On one hand, there are the potential benefits: two of the largest expenditure areas when running a data centre are the cooling and the powering of the facility, and both could be significantly reduced in outer space. The sun would provide an immense, renewable power source, while stacks are cooled by space’s chilly vacuum. But the underlying factor here is engineering.
Whilst there is an industry move towards taking advantage of ‘free’ cooling energy, the same end result can be achieved by ensuring a facility is running optimally. The management and development of data centres revolves around balancing the needs of the business with the practicality of the project. Therefore, concern is raised when additional and unnecessary risk (not to mention incredibly complex logistical considerations) are added into this equation.
Increasingly, we are seeing businesses utilising VR technologies in order to manage critical infrastructure. In this context, this would allows engineers to effectively optimise the sites without the need for a spacewalk, performing critical maintenance work from the comfort of earth using VR goggles.
Whilst taking drastic steps to reduce the overall running cost of facilities can seem attractive, it increases the likelihood of a worst-case scenario such as a hardware failure. This could be catastrophic, and hugely disrupt the running of the business. So when the likelihood of failure is higher, understanding the margins within which a data centre can operate safely is more important than ever.
The Space Simulation
Advances in simulation technology have revolutionised design and testing across multiple industries. For example astronautical engineers wouldn’t dream of building a shuttle prototype without comprehensively simulating all aspects of the proposed system’s performance. In the case of data centres, especially those built in extreme environments, the mindset should be no different. Data centre managers should use a combination of the latest simulation techniques such as power system simulation (PSS), computational fluid dynamics (CFD) and 3D modelling to effectively predict the impact that both minor and major changes can have on the facility.
Before shooting off into space, simulation allows us to consider alternative, less extreme courses of action. For example, through testing the state of a current data centre, we may discover that the hot air recirculates causing the facility to waste energy through over-cooling. This can be eliminated quite simply, through trialing a variety of optimisation techniques in a safe virtual environment. The ability to adopt this trial and error approach with multiple creative and innovative ideas can be a highly effective way to mitigate the risks associated with data centre design.
Whether managing a facility on the outskirts of Mars or Macclesfield, we have the ability to consider all of the options available to us before making a decision. While the former may seem attractive to the daring futurists among data centre professionals, it has never been easier to explore convenient and economically viable ways to optimise existing facilities.