Data centre design: it isn't just about scale, it's about reach
Members of our data centre team have been attending the virtual DCD ‘Building at Scale’ conference this week. It has been interesting to see how our experience is reflected in the industry as a whole. The phenomenal growth of the sector that we have seen over the last few years seems set to continue. According to CBRE, demand from cloud providers has accounted for 80% of take-up of new data centre space in the FLAP markets over the last year, and those same companies are looking to expand into new cities across Europe.
Covid-19 is setting new challenges for the technology industry. Remote working, the explosive growth of Zoom and Microsoft Teams video conferencing, and the rise of video streaming are just some of the factors to which we can all relate. This is creating additional demand at a time when data centres are having to cope with reduced staffing levels, and construction projects are being slowed by government-mandated shutdowns and social distancing measures.
Two buzzwords that continue to float around the industry are Hyperscale and Edge. Hyphen has been involved in designing and delivering several hyperscale data centres over the last few years, for some of the largest players in the industry – many of which will become operational in 2020. As a result, we are very familiar with the high-speed, high-density demands of this market.
Edge (devices processing data close to their end-users) is not a new concept, but as cloud platforms experience massive growth, clustered in both core and remote locations around the world, the performance of real-time critical data suffers in Tier 2 cities – those with large populations, but not close to traditional data hubs.
If you were to ask ten people what an Edge data centre is, chances are you’ll get ten different answers. From well-connected colocation data centres built on the outskirts of cities to micro-data centres built into shipping containers, connected to 5G masts, there are many different approaches that will compete to provide the low-latency data processing required to support IoT, self-driving cars, Augmented Reality and other AI-driven services. Edge will not take over from Hyperscale, but will complement it. It doesn’t matter if your photographs from last year’s holiday (remember those?) take a few extra seconds to download, but you’ll soon get frustrated if an important business meeting is blighted by dropped audio and pixelated video.
Hyperscale facilities offer highly secure, highly resilient and highly efficient facilities, maintained by exceptionally skilled staff for the centralised storage and processing of data. These facilities need a lot of space and power – preferably from renewable sources, create noise and (occasionally) exhaust, and are ideally located away from population centres.
Edge data centres, by contrast, can provide the low-latency, processing, and data caching close to populations. This benefits the service providers by reducing the amount and cost of data transferred between hyperscale facilities and end-users, whilst consumers benefit from faster and more reliable services. The downside of Edge is the lower efficiencies inherent in deploying, operating, and maintaining multiple smaller facilities.
Whether it’s re-purposing empty office space or shops, or deploying modular data centres on an industrial estate, these projects can benefit from architects who are familiar with the needs of DC operators, understand the technology and can adapt “standard” designs to different situations. Hyphen can help from site selection, due diligence and test-fits through the design and permitting process and during construction across Europe and beyond.
Over the last few years, we have helped to deliver ever larger data centres to satisfy cloud providers’ insatiable growth. Perhaps in the future ‘building at scale’ will also refer to the deployment of many smaller DCs to support growth at the network edge, as well as at its core.