Summer heatwaves - a cause of celebration for sun worshipers and tourism.
But hot summers also present a hidden potential threat for any organisation's infrastructure. As temperatures rise, so too does the pressure for on premises data server rooms and data centre infrastructure, which if not sufficiently cooled, can fail. While people think of cloud computing and the internet as something ethereal, it is a physical infrastructure that needs constant monitoring and maintenance.
Sudden outages disrupt customers and could potentially be harmful to an organisation’s reputation. British Airways learned this in 2017 when a worldwide IT failure of their systems lead to flight cancellations and delays over a bank holiday weekend. In one outage, 400 flights were cancelled and 75,000 passengers. Initially, the data centre failure was attributed by the CEO of British Airways to a power surge and a back-up failure, but speculation points to overheating as the cause of this massive IT failure. The British Airways data centres were constructed in the 1980s and were not built with today’s large-scale internet data usage in mind. As a result, the building was not equipped to deal with the excessive ambient heat generated on that sunny day combined with the scale of data usage. Thus, the equipment shut down to protect itself from the overheating. This is not an isolated incident, in 2013, overheating shut down Microsoft’s data centre for a whopping 16 hours.
Data centres need to run efficiently as well as sustainably. EkkoSense ran a study on 128 localised data centres in the UK and found 78 per cent of them were running too hot. Data centres that are running too hot are, potentially, consuming too much energy or not using the energy efficiently, so it is a cause for concern. This ambient overheating contributes to the likelihood of a complete shutdown of servers, impacting users around the world, potentially leading to data or equipment loss.
Where organisations have built data centres or server rooms on premises, there may be a lack of design consideration around the physical power consumption and cooling requirements to meet current and future demands as a business grows. This lack of virtual headroom, coupled with a sudden spike in ambient temperatures, could lead to connectivity issues or complete systems loss.
Now, with the combination of a dramatic uptake of digital services due to the pandemic and increasing temperatures, how long can your organisation tolerate the risk of an outage to a critical platform or customer service, leading to both financial and reputational damage, due to warm weather and badly designed infrastructure?
How to avoid a summer meltdown
- Outsourcing or Co-locating your critical systems and infrastructures. This will not only reduce the demands on your own IT staff to maintain uptime, but free them up to optimise and focus on business-critical priorities.
- Understand how data centres are cooled: If you have already outsourced or are considering moving systems to a data centre, understand not only how they maintain the optimum data centre operating temperature, but what backup procedures are in place should temperatures spike. More importantly, establish how they do this in a sustainable way.
For example, these are the steps we take:
- Smart Cooling: Interxion offers smart cooling islands which are fully automated and operate efficiently. During cold days we take advantage of Irish climate with free cooling unit support, where we use naturally cooled air keeps data centres at their correct temperature. This is both very economical and have a minimum impact on our carbon footprint. When we see temperatures rise sharply, our DX (Direct Expansion) chillers will support or take-over the cooling load. In DX cooling, a refrigerant coil is placed directly in the supply air stream. As the refrigerant evaporates and expands, it removes heat energy, lowering the temperature of the supply air stream.
- Closed water loops: Interxion datacentres operate a closed chilled water loop system ensuring business continuity even in hottest conditions, and are not impacted by water restrictions put in to place due to heatwaves
- Back up procedures: All our data centre facilities operate with N+1 redundancy so that if there is an issue, we have a fall-back system for uninterrupted support. Put simply, we have back-ups in place to ensure continuous uptime
With ever-changing and increasing digital demands placed on an organisations’ online services coupled with the legal and reputational impact of any service outage, why risk keeping on-premise
servers and data centres that cannot scale to meet sudden spikes in temperature and demand due to a nice summer?