Non-stop business

By Paul Tyrer

It’s difficult to envisage any business that could function effectively without IT in this day and age. In fact, anecdotal evidence suggests that even when businesses are failing or unprofitable, they continue to invest in information technology. In recent years this increasing dependency has been fuelled by great leaps forward in technology and a buyers’ market.

Today, most would agree that the best place to house new generations of compact and powerful IT servers is in dedicated computer rooms and data centers. Irrespective of whether you choose to build your own facilities, outsource your IT infrastructure or elect to have your IT supplied as a service, a key objective will be to ensure continuity of services.

Until recently, IT availability has been pursued at almost any cost, with highly resilient infrastructure designs the order of the day to ensure power and cooling for IT irrespective of the mains power supply condition. But escalating energy costs coupled with concerns about the environmental impact of increasing numbers of data centers has stimulated action from the EU in the form of the Code of Practice for Data Center Efficiency. As a result, the consultants responsible for data center design and build and the manufacturers that supply the equipment to power and cool these facilities have started to give primacy to the need to ensure energy efficiency.

This has led to a number of changes to the uninterruptible power supply (UPS), the piece of equipment that protects IT equipment against the loss of the mains supply. In the first instance, designers have sought to more closely size the UPS more exactly to the load being protected. Over-sizing of the UPS not only has consequences for capital costs, but also drives up operating expenses such as energy and service costs.

The introduction of scalable data center physical infrastructure is therefore a key enabling technology for reduced total cost of ownership. By taking a building block approach to UPS protection, modular architecture enables additional power modules and battery cabinets to be deployed as a ‘pay-as-you-grow' solution for growing IT estates. The system can be scaled back too: virtualization enables organizations to reduce the number of physical servers it requires to run many of the applications required for day-to-day activities. Scaling down physical infrastructure also affords the opportunity to maximize potential energy savings and therefore reduce cost and environmental impact.

At the same time, modularity brings benefits by enabling service parts to be kept onsite and by reducing the complexity of maintenance. Mean time-to-repair (MTTR) together with mean time-between-failure (MTBF) are the measure of availability. By making important components such as power modules user-replaceable, time-to-repair can be decreased thereby improving the availability of the power protection plant. A major cause of UPS failures is as a result of service visits and simplifying maintenance routines plays an important part in reducing this hazard.

The data center sector is in the process of adopting the power usage effectiveness (PUE) metric as a gauge of electrical efficiency. The measure is based on the proportion of facility energy that is used to power the IT equipment hosted in the data center. Ensuring that electrical infrastructure is scaled according to IT requirements is key to achieving lowered PUE, evidence of a well-managed data center.

The introduction of UPS that are both high efficiency and that achieve high levels of efficiency at low IT load conditions can be an aid to improving PUE by increasing the amount of energy available for powering IT. High efficiency UPS can also provide some overhead for growth without the user being penalized through heavy operating costs.

UPS vendors are taking a technological approach to solve the problems caused by the proliferation of technology. Current generation UPS provide a range of topologies to ensure availability and while the static UPS remains the staple for modern data center designs, the benefits of modularity, scalability and serviceability are helping ensure the fitness of the facilities that are at the heart of today's IT services.

Biography

Paul Tyrer is Vice President of UK, Ireland, Nordics & Baltics at APC by Schneider Electric. Tyrer joined APC in 1994 and has since held a variety of management roles. He has served the company as country VP since its purchase in 2007 by Schneider Electric.