Global businesses have successfully re-engineered business systems and their underlying IT infrastructures to meet the challenge of delivering 24/7 service availability. In the face of growing fears over climate change, companies are now addressing the huge task of re-orienting their business operations to reduce their carbon footprint – specifically their IT system energy demands. But as companies face up to fulfilling their corporate and social responsibilities, senior executives must ensure that reducing their carbon footprint is brought into the risk management planning for their IT operations. In particular, as companies green their operations with the latest high density, lower energy demand servers, they must ensure that they don’t inadvertently compromise the resilience of their systems by asking too much of their datacentres’ existing back up power systems. The latest servers and storage equipment are placing unprecedented strain on datacentres’ standard uninterruptable power supply (UPS) and back-up generator systems – the units that maintain continuous operations in an emergency or power outage. For example existing UPS equipment, originally designed to supply 100kW of power, is now more likely to fail and disconnect itself at 80kW or less when powering new, “greener” IT equipment. The start-up power needs of new equipment can also outstrip the capacity of older datacentre power supplies – even though on paper they can run the devices. Companies that fail to address this hidden but critical mismatch in what the industry calls “power factor correction” are inadvertently undermining the capability of their datacentre to withstand power grid difficulties. Modern IT hardware relies on a “clean” and secure power supply; having no power protection exposes the hardware, and business, to an increased risk of shut down or damage due to power disturbances.

The problem stems from three factors: the huge expectations being placed on enterprise IT, the way that many data centres are owned and configured, and the gathering environmental agenda.

First, computing development has for decades been about providing more computing power and performance – at the expense of all else. As the global economy boomed, IT departments were told to add more capacity and the CIOs scaled up the infrastructure simply to keep pace with demand. The internet explosion exacerbated the situation, demanding “always on” business applications, leading to vast computing centres, with bigger servers, more storage, and powerful cooling systems that themselves required additional power .

The second factor is the manner in which many complex datacentre infrastructures are operated. Datacentres may be duplicated at multiple locations for business continuity, or are sprawling estates that have grown massively to meet demand. They may be co-located and run by third parties, maintained by outside infrastructure specialists, all of whom must support the IT department. Ordering an “about turn” to deliver green IT requires input from many specialists.

Third, the climate change debate has caused further disruption to business strategies. As critics have suggested that IT operations have similar levels of carbon emissions to aviation, the IT supply industry and its users have had to reverse their power consumption habits – while still meeting the 99.9999 per cent service expectations of the online user.

Everyone appreciates the benefits of energy efficient operations but changing the demands on complex datacentre infrastructures must be a central part of an organisation’s risk management whether it adopts a declared green IT strategy or not.

The answer is for organisations to engage with their datacentre providers and infrastructure specialists to examine the operational requirements and power demands on datacentres to identify the critical pressure points and risk factors. Enabling issues such as power factor correction have to be recognised and managed while supporting the overall IT strategy. As investment in more energy-efficient operations is approved and existing datacentres gradually become rows of ultra high density servers and storage arrays, datacentre UPS and generator protection will require regular risk assessment, performance testing and, where necessary, certification that they are still fit for purpose.

This is no phantom menace: infrastructure experts have started to see the first instances of what were initially unexplained failure of hitherto robust and proven UPS in datacentres.

The CEO that walks through a datacentre would doubtless be impressed by a state-of-the-art blade server that handles unprecedented numbers of transactions with greater energy efficiency and a smaller footprint than ever before. They might be less impressed if they knew that because the IT department didn’t go through the extra risk assessments with the datacentre owner and their infrastructure adviser, that the server’s back-up power which seemingly ticked all the insurance boxes, is now an emerging risk to their organisation. 

Back to The Green Room Homepage>

More articles:
The Green Datacentre - is it possible?