Future proofing the data centre industry

Based in London, Future Facilities was born into the emerging data centre (DC) industry in 2003. It creates software suites specifically tailored to the DC market for the design, ongoing operation and maintenance of these facilities. This can cover anything from ground up to ground-breaking, design work, new tech employment or troubleshooting and retro-fitting existing centres, while also planning new phases of IT equipment within an existing facility.

“We cover the cradle to grave of a data centre,” maintains Future Facilities COO Jonathan Leppard. “On the software side, we’re utilising simulation to predict what will happen so you don’t have to wait to react to problems. We also have an electronics software product which examines the thermal dynamics of hardware so we can simulate laptops and computers down to chip level. We also provide consultancy services to either ‘out-house’ an entire project as the third party or train people with the technology so they can bring it in-house themselves.”

Future Facilities was shortlisted for the Mission Critical Innovation Award at the 2017 DCD Awards, recognising its quest for excellence using DC simulation in virtual reality. The engineering simulations it provides allow DC owners and operators to understand how to refurbish efficiently and how new construction can be better managed in terms of capacity, energy consumption and cooling. “People will have grand plans but the reality has to hit at some point,” warns Leppard. “The beauty of simulation, replicated in many other industries, gives the ability to look at things in the round before they have to be set in stone. We build these DCs for capacity, efficiency or resilience – some will need 24/7 resilience while others may only need a portion of that. We can then start providing solutions tuned to the needs of a business rather than solutions for the performance of a facility.”

See also:

• How IoT can transform the construction industry

• Global power generation construction project market worth $4.78tn

• Global data centre construction market to reach $22.8bn by 2025

Fifteen years of the company’s expertise is built into its tools and solutions such as the 6Sigma range. “When our users are simulating DCs, it’s important that what we’ve learned is built into the defaults to quickly give them the answers they need to understand where the efficiency losses or gains might be in their design,” asserts Future Facilities Product Manager David King. “In the past decade the rules of thumb people apply have improved, but it’s the small details which our tools can provide for construction models which can help owners and operators achieve better insight into how their DC is going to perform in the future.”

The innovation of Future’s offering lies in having a software package that delivers fundamentals in line with the industry for informed decision-making. “We bring together usability and functionality with industry collaboration to provide the end user with a future-focused qualitative result,” maintains Leppard. “One of the biggest drivers is that no one can do this by themselves – this is a community effort. We’re confident we can bring together disparate silos such as IT and facilities with a tool that gets communication flowing to support better, quicker, long term decisions.”

With DCs getting bigger and calculations becoming increasingly complex, Future Facilities partners with cloud providers to give users access to multiple cores on a price per hour basis so they don’t need to buy expensive hardware in house. “Virtual reality is something we’ve been embracing so we’re able to put our users inside the model using Oculus Rift,” adds King. “Looking to the future, we’re aiming to add some of our simulation results into augmented reality to help visualise the invisible aspects of cooling systems in DCs such as the air flow. We’re also investigating ways of automating report generation and diagnostics so our users can fully understand their simulations results and get the full benefit.” Future is also planning to apply its tech to consider the effects of where DCs are located, using external modelling to look at efficiencies using air instead of air conditioning to make DC services much greener.

A pioneer in the space, Future is excited about the potential for simulation to sit at the heart of the design process for mission critical construction, and up for the challenge of promoting its integration with operational planning and process management. “We think we can provide this for the DC life cycle,” assures Leppard. “It does mean changing processes to accommodate another cycle of thought, so speed of adoption in the industry will be a challenge for us. Our software is evolving to meet the changing needs of management and monitoring. Understanding the ROI is key with something you can’t necessarily see or monitor in a physical sense. You wouldn’t necessarily know it’s working for you unless you had a major failure.”

Future’s partnership with its cloud provider ReScale (a dedicated simulation hardware provider which uses most of the major cloud providers as infrastructure) allows it to run on large compute up to 256 cores on Amazon, Microsoft and Google. “We also integrate with a providers of DCM tools such as SNT Group, Schneider, Trellis from Vertiv, Intel and Panduit,” says King. “Our model needs data that might already exist in other systems and we don’t want to duplicate that. What we do is complementary to, rather than in competition with, these major software providers.”

Impressive levels of industry adoption highlight the appetite for simulation tools, like Future’s 6SigmaDX, in the construction of colocation DCs. For example, when Compass (a colocator responsible for advanced disaster-resistant, future-proofed DCs) needed to be able to capacity plan with confidence, it chose Virtual Facility to design and commission DCs that maximise revenue and ensure service level agreements (SLAs) remain un-breached.

“We also work with hardware vendors like Airedale,” adds King, “which uses our software to design their air handling units as they are looking at how it internally sets up and configures its cooling units to get maximum benefit from outside air conditions.” Elsewhere, City Bank in the US is utilising these simulation tools for energy efficiency savings.

The team at Future is well placed to react to current trends in the construction of DCs. “We’re seeing a move from chilled water and DX units with a cooling plant outside serving air handlers inside, towards direct fresh air cooling or indirect evaporative coolers on the external wall of the DC,” notes King. “To combat many designs which would have fallen over on a hot day, the cooling is done directly through air to air exchange without the transport medium of water or glycol. It changes the dynamics of the outside of buildings because you have the same problem – to deliver enough air inside the data centre to all the IT boxes to deliver processing power, whilst on the outside you have containment issues to squeeze all the large air movement devices into a small space with changing ambient conditions and wind profiles. Therefore, it becomes just as important to simulate outside conditions, which could be potentially catastrophic when they are delivering 30% of your cooling.”

Weighing up predictions for the future of the industry, King notes an increase in the complexity of control systems for cooling, moving away from individual air handlers to system-based buildings. “We’re not sure complexity means better, but we’re certainly seeing examples where the complexity doesn’t seem to be offering obvious benefits,” he warns. “People will try and get as close to the edge of the efficiency envelope as they can by adding layers of complexity with fans that ramp up and down for a trio for conditions. We’re not convinced the efficiency gains people hope for will be delivered.”

This reckoning with reality could be overcome with AI and machine learning for control systems. Future believes the challenge will be creating the ability to learn in heterogeneous DC environments that change quickly. “Simulation has a place in that as we move into automation,” says Leppard, “but we’ll give the machine learning something to reference, prior to a DC’s launch, to bulk up the start point and then allow fine tuning with the reality that ensues.”

The ultimate goal in 2018? “We’re focused on making simulation a part of everyday processes and we’re starting to see this with corporates. Meanwhile, we’ll continue to work with research companies like ES2 and The Green Grid on performance indicators to provide tools to help the industry and give better insight to decision makers.”