Empowering excellence in data centre design

Over the years I’ve worked in several different roles as both a design consultant and as a data centre design engineer (SME) and, in doing so, gained a good insight into the difficulties and opportunities facing data centre design. My first blog seeks to ask the question ‘how can we, as the data centre industry, work with our design consultants to improve the quality of our infrastructure designs and be more innovative or pioneering?’.

At Kao Data, for example, we believe we need to innovate to remain relevant in this fast-moving industry, primarily to ensure we keep up to date with our customer’s needs and constantly drive down our energy demand.

We think part of the solution can be found in empowering our design consultants and allowing them more room to drive the design stage.

What’s the problem?

Within the data centre industry, we tend to be ‘technically intelligent clients’ in that we are generally from a technical background and understand the complexities of engineering in more detail than many other adjacent sectors. As a result, we often to tell our designers what to do rather than ask them what we should be doing, which can lead to consultant designers becoming more passive and letting themselves be led, rather than leading us – i.e. you do what the client asks!

This approach isn’t always beneficial, and speaking transparently, we don’t know everything we think we do so may be driving the design in a less effective direction. As an industry, we do this in several ways, such as determining our concept design configuration before we even engage with the design team.

We often issue very specific and technically detailed employers’ requirements, e.g. Basis of Design (BoD), project of reference (PoR) etc. One BoD I saw, for example, ran to over 600 written briefing pages and left little room for the designers to influence or improve the design.

At Kao Data, we believe this stifles innovation as designers don’t always feel they can question, challenge, or suggest new ways of doing things. We can pay consultants a lot of money to undertake the design, but then don’t give them a chance to fully earn those fees.

There are benefits to this approach…

Why do we do this? The answer is simple – to drive consistency across data centre buildings and campus’s and ensure all our high performance data centres are designed exactly as we need them. If executed correctly, this can speed up the design programme, lower costs and hence reduce overall project delivery programmes.

Design consistency also helps with equipment procurement, reduced build times, commissionability and offers many other benefits. Finally, it allows the on-boarding of new design consultancies, with little or no knowledge of “your way of doing things”, enabling them to get up to speed in a much shorter time.

How does this help with innovation?

I’ve seen, and worked on, many Basis of Design (BoD) documents and they tend to be reviewed and updated gradually, iteratively and with minor changes. Changes, improvements, and lessons learned can creep in but once the design is fixed on paper, all parties tend to see this as the ‘design bible’ that must not be radically changed. Strangely, this can also affect the client-side engineering teams who sometimes see the BoD as sacrosanct rather than a design framework.

Design Quality and staff retention

What does this do to our design consultants? Design teams from many consultancies are currently losing a lot of key talent as the engineers become bored of copying and pasting someone else’s design. Good engineers feel they are not developing their engineering skills and it becomes harder to take pride in their work when they’ve not had much influence in the design. The quality and consistency of the designs can subsequently suffer.

This is becoming a real problem for design consultancies who are struggling to recruit and retain good people which in turn hurts our industry by exacerbating the skills shortage, something which has been well documented by the Uptime Institute.

It’s not all doom and gloom though. Data centre design is still challenging, interesting and rewarding work, much more so than I’ve experienced in other adjacent sectors. However, to keep the talent in the industry, and ensure designs remain relevant and engaging, we must consider the people who produce them.

What can be done?

Whilst we, at Kao Data, have a technically advanced and highly experienced engineering team built of industry leaders, who know what we want and have our own BoD, we also want to ensure that the data centres across our high performance infrastructure platform are efficiently, competitively and innovatively designed. The following are some of the initiatives we encourage:

• Bring the designers on board as early as possible – e.g. don’t fix on a site test-fit layout before it’s gone through the wider ‘ideas mill’.
• Write a basis of design (BoD) document as a ‘framework for the design’ and only be specific about key requirements - leave room for the designer to innovate and hence take both ownership and pride in the design.
• Be clear about what you want, but also why you want it. For example, summarise your typical SLA requirements, business drivers, insurance requirements etc. This will help the design team better understand your design motivations and lead to the design you actually need.
• Workshop ideas collaboratively throughout the design process. Give your designers the opportunity to bring YOU ideas and don’t be too bound by your own BoD. Encourage the concept that “no idea is a bad idea”.
• Allow time within the design programme for innovation. Developing new ideas often takes a little more time. If the programme only allows ‘just enough’ time to “crash out” the specs and drawings, little will change.
• Recognise that your designers work for many data centre enterprises and can bring this variety of experience, ideas and lessons learned into your project design. Recognise that you may be in a “bubble” and there are many ways of doing things.
• Be genuinely open to ideas, suggestions, and change. Just because it worked on one project, doesn’t mean it will work on all projects or that new ideas won’t work equally as well – or even better.

Final thoughts

If stakeholders within the industry simply ‘rinse and repeat’ our designs, we will be in danger of stagnating. For example, in the past decade, the industry has moved from poorly arranged data halls to ‘hot and cold aisles’, to ‘contained hot aisles’. Each of these innovations has helped drive up power densities and improved data centre efficiency. I believe we must continue to encourage new ideas from our design teams to enable us to quickly move forward into a world of high density processing, liquid cooling, HPC, AI, and quantum computing.

It is also critically important we ensure our designers are involved in all key stages of the design and are engaged and motivated stakeholders. It is, unfortunately, becoming increasingly difficult for consultancies to recruit and retain new staff, engineers are moving roles regularly (which impacts project continuity), which is in turn putting a real strain on the industry, and the quality of the designs being produced.