Published in the Winter Data Centre Management Magazine, our Head of Design; Stephen Lorimer looks at resilience and how the industry mindset needs to change.
A common definition of ‘resilience’ is ‘the capacity to recover quickly from difficulties’. When applied to the data centre sector it is more commonly accepted as: the ability of an IT infrastructure to continue to operate, for example following an issue such as power outage, equipment failure or human error.
There is a general misconception that all data centres should be highly resilient. In fact, I lose count of the times that customers have started initial meetings by requesting a “Tier III or Tier IV facility and, above all, absolute protection against any data loss.” Quite often by the end of the initial engagement they realise that they are already achieving the redundancy they need within their IT layer and can normally operate safely with a lower resilience classification.
Historically organisations have often made the mistake of designing highly resilient data centres without properly considering why and whether or not they actually need them to be highly resilient. As a result they have ended up with, at best, facilities that have been both expensive to construct and continue to be operationally expensive, and unnecessarily complex, to operate.
At the heart of this problem is the fact that decisions about the resilience of supporting M&E infrastructure are often made without any proper consideration about what level of availability the IT service is actually required to deliver. To do this involves taking a step back and looking at the wider IT strategy – an approach that is endorsed in the EU Code of Conduct which clearly states that organisations should “deploy resilience in line with requirements.’
This failure in approach can happen for a number of reasons, but the most common one we see is that organisations do not engage with a specialist. Whilst in-house teams are often extremely competent, data centre design is rarely their ‘core skill’ so the end result may not always meet and sometimes contradicts the company’s IT objectives.
For those occasions where design and build is the best option we, as an industry, need to put more of a focus on ensuring that data centres are ‘designed for operation’ and the team responsible for maintaining and running the facility is engaged from the outset. As an organisation we encourage different stakeholders, to be part of the process from the outset, as we feel that this delivers the best results. This early engagement is key as not having all the stakeholders involved may mean that not all impacts are properly considered and addressed as part of the design…