By Honeywell Senior Data Centre Director, Alpesh Saraiya
Data centres play a critical role in keeping the global economy productive. Demand for data storage and processing has become insatiable worldwide, which makes it more challenging than ever to operate and scale these facilities efficiently. Managers also face mounting pressure to make their operations more energy efficient. Data centres consume about 3% of the world’s electricity – more than most countries – and produce 2% of global carbon emissions, which is about the same as the entire airline industry.
Given the exponentially increasing demand, managers are often forced to do more with less, while at the same dealing with tougher internal environmental, social and governance (ESG) directives and more stringent regulatory landscapes. When preparing for 2023 and ahead, operators should factor in three significant trends: continued pressure to cut operating costs, increasing demand for more sustainable facilities, and a growing shortage of talent interested in and qualified for managing data centres.
Rapid scaling with a sharp focus on managing OpEx
To meet the unrelenting demand, both hyperscale and colocation data centre operators have aggressively acquired smaller firms – but this practice has created as many problems as it has solved. For one, it creates a pastiche of ‘snowflake’ designs – no two are exactly alike – which means major headaches for integrators and heftier operating expenses (OpEx) for owners.
Blueprinting data centre designs to achieve commonality across facilities has thus become a key strategy for ‘doing more with less.’ While many see it as a critical step toward developing and implementing global design standards, operators must still comply with local and national building codes, financial accounting laws and security regulations.
Many data centre managers are also streamlining their operations with well-defined, purpose-built workflows and operational management tools to further reduce OpEx while still protecting uptime. To this end, some are installing cross-domain, site-level monitoring and management platforms to automate as many tasks as possible, thereby easing workloads and reducing the chances of human error, which caused major outages over the last three years at 40% of organisations surveyed by the Uptime Institute.
Tougher internal and external sustainability mandates
Data centres face mounting pressure from governments, clients and stockholders to become more sustainable and energy efficient. A sustainability strategy is no longer simply a ‘nice to have’ item, in the future, it may determine whether an operator succeeds or fails. With financial firms at either end of the transaction – as both clients and providers of capital – operators will face an additional hurdle of expectations when seeking to fund future projects, particularly as pressure increases on private equity and real estate investors to make greener investments. Further, even clients shopping for data centre colocation service providers are now examining ESG profiles to account for upstream Scope 2 and 3 carbon footprint.
As data centre operators struggle to scale up across different geographies, they face a range of ever-stricter local and national regulatory landscapes. Many of these include increasingly rigorous sustainability and ESG financial reporting standards that will phase in over this decade.
Governments including Ireland, the Netherlands and Singapore are requiring owners and operators to submit a detailed sustainability plan before granting them approval to build a new facility or expand an existing one. Singapore, in fact, imposed a moratorium on new data centre projects in 2019, and kept it in place until January 2022. Applicants for new projects now must explain how they will meet tough new standards enacted to protect the nation’s land, water and renewable energy resources.
Increasingly, governments are expecting data centres to measure and disclose their carbon footprint and demonstrate progress toward reductions. There are numerous ways to reduce carbon emissions – no one size fits all – but cost and new technologies usually factor into the equation. Among these, operators are evaluating a variety of energy optimisation techniques, from control loop optimisation to liquid cooling options, especially as high performance computing (HPC) and AI/ML applications become more ubiquitous and more demanding in heat dissipation requirements. Air cooling systems simply can’t keep up with the cooling needs of continually evolving, higher-density racks for these next gen workloads.
Liquid cooling leverages the higher thermal transfer properties of special fluids, providing as much as 3,000 times the efficiency of air cooling. With more and more businesses integrating HPC applications driven by AI – which require tremendous computing power – operators are realising that the time has arrived to seriously incorporate liquid cooling in their architecture and roadmap.
The human element: a looming talent shortage
Not least among data centre challenges is a widening skills gap and the ongoing ‘great resignation.’ Some see this as a result of the industry’s ineffective efforts to actively recruit and retain talent from the vocational schools over the last two decades. The industry is also dealing with an aging workforce of subject matter experts – those qualified to teach entry-level employees – many of whom will retire within the next 10 years. Yet Gen Z workers who have the skills and aptitude to pursue such a career are not seeing careers in data centres as an attractive option. Nevertheless, there are promising initiatives in the industry to source candidates from a pool of disciplined and well-qualified military veterans.
As operators develop plans for 2023, they should be considering strategies for scaling up intelligently, reining in OpEx and prioritising sustainability efforts. They should also take a hard look at how they can make the profession – and their facilities – more attractive to the next generation of talent.