2022 Predictions: Digitising data centre design and build
Data centres have become the very heart of the digital economy, and critical to our ever more digitised way of life. As we adapt to a new and hybrid world, greater innovation will be necessary to help overcome many of the remaining challenges, including the need for increased sustainability, more efficient use of energy, and for our industry to meet accelerated demands for capacity. Let’s take a closer look at five trends that are influencing the direction of data centres.
Digital design tools speed development
I expect to see greater innovation in the digitisation of data centre design and build. One of the top challenges customers are experiencing is the need to meet demands for new data centre capacity. To help address this challenge, new software tools are emerging that speed up the design and construction of data centres. Schneider Electric’s partner ETAP produces software (essentially a digital twins tool) that allows designers to model the electrical powertrain for availability, efficiency, and sustainability. Another company, in which Schneider Electric has a stake, is RIB, which develops construction management software.
Traditional computer-aided design (CAD) platforms have long allowed users to design the layout of a facility, however, the use of ETAP’s software allows detailed modelling of the powertrain while RIB’s enables time and cost modelling. Although CAD tools have been familiar for many years, the ability to model the powertrain is new. End-users can now choose or substitute components and subsystems based on their environmental impact or energy efficiency - evaluating the effects on technical performance and pricing via digital twins before committing to physical prototypes.
The 6G effect
Fifth generation networks have been expected to make an impact for some time, but the fast millimetre wave 5G variant has been slow to materialise. 5G is, however, beginning to make an impact in open spaces with few physical barriers such as stadiums, airports, and shipyards. The problem remains that a killer application to drive the need for mass adoption has yet to materialise.
An exciting prospect is 6G networks, which could offer life and experience changing functionality. 6G operates at THz frequencies and has access speeds of 1Tbps, which will deliver near ‘air latency’. Whereas high band 5G hits speeds around 500Mbps, with air latency aimed at 8-12ms. Potential use cases for 6G include embedded technology for controlling artificial limbs (prosthetics) through wireless Brain-Computer Interactions (BCI), which is an incredible prospect! In the 6G world, people could interact with their environment and other people using devices that could be held, worn, or implanted.
6G networks also have the potential to eliminate traditional base station and antenna networks because their high frequencies need a ubiquitous mesh network where everything around you has an antenna function. In theory everything that powers up will have a built in antenna function and become part of this new ‘antenna free’ network. While the network architecture may change with 6G, the computing capacity will need to grow, so placement at the edge will become even more crucial.
Energy concerns at the edge
Adoption of edge infrastructure will also continue to grow. However, energy efficiency will become a critical factor, with customers demanding that edge deployments match the capabilities of larger data centres in terms of resilience, efficiency, and sustainability. Edge deployments may be smaller than traditional facilities, but the scale and volume at which the infrastructure is likely to be deployed demands its environmental impact be minimised.
Building a sustainable edge at scale requires greater attention when selecting components, during the design and deployment stages, and use of comprehensive management systems to drive operational efficiency. Cooling will remain an essential part of the efficiency requirements, but the challenges presented by edge deployments, especially those in unmanned environments, will require innovative approaches in terms of technology and topology.
Air cooling is often unsuitable for edge deployments, which are frequently located in urbanised and harsh locations where dust and other contaminants abound. Blowing such material around an unmanned or remote edge data centre is far from ideal, and even if filters were attached, the task of frequent replacement and servicing remains a key challenge – especially where cost and circularity are concerned.
With sealed and unmanned edge data centres, therefore, liquid cooling will be required, although it is not yet clear what sort of topology will be best suited. As such, new liquid cooled architectures may emerge for the edge at scale. Whether that involves direct-to-chip liquid cooling or chassis-based immersive cooling is yet to be seen.
Standardised metrics for sustainability
The circular economy - the ability to reduce, reuse, and recycle technologies deployed at the edge - will be an important consideration in 2022 and beyond. However, another area growing in importance is the need for standardised sustainability metrics. Today, there are a plethora of metrics from which to choose, with data centre operators each reporting their own preferred measurements. However, I believe there is a need to measure sustainable progress in a consistent and organised way.
According to the Uptime Institute, IT and Power consumption, and Power Usage Effectiveness (PUE) remain the top sustainability metrics tracked across the industry. While PUE has long been an excellent marker of efficiency, we must also agree on metrics for the other categories of environmental sustainability – greenhouse gas emissions, water use, waste, and biodiversity.
Going forward, I believe sustainability metrics within the industry must evolve and become more standardised. This effort can leverage business processes, like GAAP balance sheets and income statements, to provide a ledger where each company can state the results using established rules and units of measurement. An approach such as this ensures comprehensive reporting that is universally understood and provides a baseline to measure success. Further, it makes it possible to compare sustainability results with other companies.
At Schneider Electric, we know that not all companies are at the same stage of their sustainability journey, which is why we recommend a framework for Beginning, Advanced, and Leading. Beginning companies will report on energy use, Greenhouse Gas Emissions (GHG), and water utilisation. The 11 metrics for this level are a mix of measured values like GHG emissions in mtCO2e and ratios like Carbon usage effectiveness (CUE) in mtCO2e/kWh. ‘Advanced’ metrics bring in the ‘waste’ category and ‘Leading’ metrics will include a category for land and biodiversity.
Data centre functions become services
Data Centre as a Service (DCaaS) offerings are beginning to gain popularity. The trend is enabled by standardising power, cooling, IT and storage in data centres to offer the same user experience and data access from everywhere. Companies like Microsoft and Amazon have already started offering such services with their Azure and Outposts initiatives, extending versions of their cloud architecture into the edge environment where customers can pay a monthly service fee for their capacity.
Many traditional IT companies such as Dell and HPE have positioned themselves as IT advisors to help companies design and run business application or workloads in the cloud (consulting services, engineering, integration and management), rather than as IT hardware and software suppliers, so one might predict that DCaaS will continue to gain traction.
Overall, I believe data centre capacity will continue to grow at both the core and edge driven by digital acceleration and enabled by high capacity networking 4/5/6G and WiFi 6. Model based software will be leveraged to bring efficient, resilient, and sustainable data centre capacity online faster, which is great timing, as we are at the precipice of edge being deployed at scale.