Our world is rapidly shifting, due to advances and take-up in technology, COVID impacts, and changes to economies and political environments. If there is one thing the business world has learned from recent experience, it’s that to effectively handle these constant and rapid changes, flexibility and adaptability are crucial to business survival.
This rings especially true for the data centre sector, which continues to see significant shifts, from the type of occupants (enterprise to cloud) and what their specific and varied requirements are, to huge changes around sustainability goals for data centre operations as owners and tenants focus on decarbonisation efforts to transition their businesses towards net zero carbon. Never before has data centre design been more led by tenants’ demands, and owners and operators must adapt to stay in the game.
With data centre equipment, including thousands of servers, cooling systems, storage drives, and network devices, representing up to 4 per cent of Australia’s total energy consumption and approximately 10 per cent of the world’s energy, there is a huge drive across the sector to decrease carbon footprint. This is being achieved through more efficient and cost-effective energy consumption, and ensuring energy is from renewable sources.
Leading the way are the likes of Microsoft, Google, Apple and Amazon, who have all committed to putting no excess carbon into the atmosphere. Within their targets is specific mention of the role their data centres will play as part of their supply chain, to help them manage their transition to net zero or, in some cases, carbon negative.
For example, Google is aiming to run all of its data centres and offices on carbon-free energy by 2030. As part of Microsoft’s ‘moonshot’ commitment to be carbon negative by 2030 and, by 2050, to remove from the environment all the carbon the company has emitted since it was founded in 1975, it included the requirement to disclose greenhouse gas emissions in its Supplier Code of Conduct.
This drive for greater sustainability, combined with the ongoing and simultaneous shift from enterprise to cloud, is resulting in a vast array of, sometimes conflicting, needs and requirements between different data centre tenants: from changing redundancy, reliability and capacity requirements, to changing environmental requirements, such as temperature and humidity conditions in the data hall, which can have major energy saving and cost reducing implications.
Historically, equipment in data centres was considered very sensitive, with strict temperature and humidity requirements. But as equipment design improves and its performance data is better understood, many tenants are finding that to run optimally (and meet sustainability goals), they can relax temperature requirements and run equipment at greatly elevated temperatures (say 30 degrees) and with wider humidity bands, decreasing the amount of cooling required.
While increasing numbers of tenants are taking this approach, there are still others who keep the dial at the more conservative level (24 degrees plus or minus one degree). Herein lies one of the challenges for data centre owners and operators:
Stranded plant refers to the situation where spare power and cooling capacity cannot be accessed because of the load distribution and configuration of the plant. A good flexible design, combined with a well-considered staging plan, will minimise the risk of creating stranded plant.
The issue of capacity not needed for several years and the associated energy inefficiency of operating plant at low load is usually managed through deferred installation. This requires a smart design that allows staging of the construction where future stages can be safely constructed and without impacting the ongoing operations of the data centre.
By prioritising flexibility and adaptability, data centre owners and operators are better placed to not just meet tenants’ varied needs, be it for enterprise, or cloud, or both, but also to realise cost efficiencies, maximise the versatility of infrastructure, minimise underutilised equipment and avoid adverse impacts on sustainability targets.
At the STT Thailand data centre, Aurecon’s team embraced flexibility and adaptability to create a facility that sets a new benchmark for design, build and operations of data centres.
Designed to operate at two different power densities depending on the tenant, it was initially configured at a density to support enterprise customers, however the same infrastructure can be reconfigured to provide the higher power density required by a cloud tenant. The infrastructure design meets the differing redundancy/reliability requirements and increased capacity, while minimising costs and wasted infrastructure.
Adaptability and flexibility also extends to ensuring cooling and power systems capitalise on opportunities to reduce energy, without compromising the ability to meet the requirements of a more traditional tenant.
Flexible cooling systems can be configured to reap all the energy benefits of a tenant who accepts a much wider band for their space temperature, but at the same time can be configured to meet the specific requirements of other tenants without having to compromise. It could be designed to deliver 24 degree air or 30 degree air if that’s all that was needed, without having to compromise on energy efficiency.
Flexible power systems incorporate built in ‘blocks’ (standby power, transformer, switchboards, and UPS when required), meaning that the power systems can be configured in multiple ways to respond to individual tenant needs. These versatile blocks enable power to be configured to meet the needs of a more conservative “2N” tenant, for example, with fully redundant primary and backup systems, and then easily reconfigured to meet the differing configurations demanded of cloud providers, for example three primary systems, with one backup, which creates an extra 50 per cent of useable capacity without installing any more kit. The blocks can also be removed when not required such as the UPS block for tenants with batteries in rack.
To achieve this flexibility, it must be integrated into the design from the outset. Technology selection, site location, systems and services must all be included in the planning and designed with interfaces and connections to cater for future growth, loads and changing needs.
It’s also crucial to design detailed upgrade strategies that enable modular plant expansions and space to be added. Don’t compromise on building in flexibility during the design phase; allow for different power density zones, plan for different plant configurations, and provide flexibility for white space rack layouts.
No longer is data centre design a one-size-fits-all approach, and the ability to cater for the varying needs of tenants, without pulling equipment out at great cost and waste, is a business must-have. Increasing demands from tenants to meet their sustainability goals is also putting pressure on designers, owners and operators to ensure data centres are exploiting every opportunity to decrease carbon emissions.
Those who ignore flexibility, adaptability and sustainability in design risk ruling themselves out of a significant part of the market. However, those who embrace these imperatives will keep the door open to opportunity and stay competitive in the data centre game now and into the ever-shifting future.
Phil Motteram leads the design of large and complex building projects for Aurecon clients across Victoria and South Australia, specialising in data centres and essential social infrastructure. Phil previously led Aurecon’s data centre business in Asia, leveraging over 20 years’ experience in the design of highly reliable, energy efficient data centres.
Please change your browser to one of the options below to improve your experience.