Monuments, dark arts and urban myths

Data centre infrastructure has naturally evolved in response to new technology, but thankfully the UK data centre market has also become much more open minded about what actually makes a good starting point for a data centre design. Equally, there is a growing realisation that by starting with a lack of pre-conceptions, you can actually build to achieve improved efficiencies.

Not only have we seen a move away from an overly simplified approach to design that was common in the early days of the data centre, when the focus was simply on finding a big enough room with enough power and air conditioning to support the networking systems prevalent at the time. It is also far less likely that the over-complicated and impractical designs of the Dot.Com boom will ever be replicated again.

With new attitudes towards spending taken into account, I expect all will agree with me that there is no value in designing a data centre facility that simply serves as a monument to the architect. Cost-effective solutions to real business problems are what is being demanded. Today’s challenge therefore is how to continue the process of data centre design evolution in the face of dynamic changes in IT - whilst thinking ahead about the potential longer term impact on the data centre.

If you consider the existing data centre conundrums when identifying potential sites for development, for example, it’s no surprise that the cost of delivering power to the site is a major issue. Optimised cooling, renewable energy sources and new build techniques, that offer improved operational efficiencies, are all being closely examined to reduce the need for power but the skyrocketing cost of power itself cannot be ignored.

In the same way, escalating real estate prices can often mean that a city centre facility is increasingly out of the question. Whilst a suburban location might offer more realistically priced land, the current trend towards out of town data centre industrial estates may just re-locate the cost problem in the long term. Plus, in this example and with multiple providers on a single site, who will then manage the allocation of services without favour to ensure all tenants have sufficient provision?

Data centres absolutely do have a ‘shelf life’ and many companies are outgrowing their facilities faster than they can start a new build or undertake refurbishment. As a result, I believe that many companies will now look to consolidate their distributed, outdated IT infrastructures into one centralised facility; driven perhaps by an aim to reduce costs and increase operational efficiencies. However, even when taking this route, the development of a larger data centre to meet increasing server and rack space requirements will in turn increase an organisation’s carbon footprint. And the time and cost of the relocation will no doubt be vastly underestimated. There is a very real risk versus cost question that must be investigated, and answered, first.

Combine these issues with the growing popularity of cloud computing; the need for heightened security; the call for mobile devices to be actively controlled in order to protect intellectual property (IP) and the exponential growth in Big Data, for example, and it’s no surprise that flexibility, scalability and adaptability are now top of the agenda for the data centre.

Going forward, it’s imperative that instead of being just a follower of IT fashion, progressive data centre design should be focused on solutions that follow best practice. Design has to be business-led by data centre experts that can translate technological advances into operational benefits that can be understood and endorsed by the board. It should not be about creating a facility that is under-used, that is too expensive to cool and requires a large budget to maintain.

Modern data centre design should be all about:
Offering value: the data centre can no longer afford to be seen as a cost centre with soaring costs. Minimal improvements in services or performance should be replaced by a commitment to innovation and a direct link to business objectives.

Improving specification: over specification simply means higher energy consumption by power and cooling systems, increased staffing levels and more maintenance. Specifying only essential infrastructure will automatically reduce all these factors.

Considering new cooling options: Alternatives to computer room air conditioning systems (CRACS) are growing in number. Also consider running the data centre ambient temperature at a slightly higher level in order to reduce energy use. Implement a ‘lights out’ policy where possible.

Improving lifecycle management: using the latest technology is one of the best way to improve energy efficiency and reduce power consumption. Refresh systems regularly to support operational goals.

Making it modular: A flexible approach to capacity is needed to keep pace with peaks and troughs in demand. Consider individual modules, pods or completely containerised data centres which can be more readily swapped in or out to meet business requirements.

Increasing the link between BIM and DCIM: convergence between Building Information Management (BIM) tools and Data Centre Information Management (DCIM) should be exploited to the full. Combining building modeling with IT equipment modeling can help you to run ‘what if’ scenarios to identify where change is needed, for example. It can also enable you to combine asset management with maintenance to minimise downtime.
Planning ahead: effective lifecycle management is important but so is minimising the cost of closing down a data centre at the end of its usefulness. Avoid housing obsolete or under-performing IT in unnecessarily large rooms with high overhead costs by planning for migration, understanding what residual values exist for IT and how the building itself can be re-purposed or sold off.

In emerging markets, particularly those seeking to provide data growth solutions, there are many valuable lessons to be learnt from the UK and Europe. Delivering a cost effective solution today that is also a ‘best fit’ for future environments and business climates is achievable however, especially if flexibility remains at the heart of the final design.