In implementing and operating every data centre, from the smallest to the largest, IT professionals face a wide and ever-expanding range of challenges. Putting in place a coherent physical and power infrastructure, made up of elements that work harmoniously together is, says Paul Ryan, UK segment sales manager for Eaton, the key to addressing these challenges successfully
Today’s data centres range in size from small installations serving the needs of an individual business or group of users to the enormous sites that support the likes of Google and Amazon. It would be tempting to think that the requirements of those who operate these very different installations would be equally diverse, but they are not. The key business demands for every data centre are, in fact, very similar.
In essential terms, every data centre operator is constantly looking for ways to improve energy efficiency without sacrificing uptime, and for flexible solutions with inherent scalability that will help to keep implementation times to a minimum. For data centres that offer third-party services, short implementation times mean faster time to market and earlier generation of revenue. For those offering in-house services, the key benefit is faster time to availability of IT resources.
Meeting these requirements is far from easy; in fact, it could be argued convincingly that it is becoming more and more difficult as expectations rise. Consider energy efficiency, for example – levels that were considered to be excellent just a few years ago are now seen as average at best. Similarly, days and even weeks are continually being shaved from project implementation schedules. And there’s another factor that’s making life even more difficult in many organisations – the dwindling availability of engineering resources as cost-control measures lead to staffing levels being cut to the bone.
Though these challenges might appear daunting, there is an effective and accessible way of addressing them, once it is understood that it is the infrastructure of a data centre – including the power chain – that largely determines its efficiency, resilience and flexibility. In other words, if the infrastructure is optimised, the challenges will be resolved. That’s an easy enough statement to make, but what exactly is ‘optimised infrastructure’? The simple answer is that it’s infrastructure made up of elements that not only offer excellent performance and flexibility in their own right, but also work together efficiently and reliably.
Now it’s true that infrastructure meeting this description could be configured using elements from multiple suppliers, but this would require a great deal of care, much time and effort in scrutinising suppliers’ data and a high level of engineering knowhow. Even then there might be lingering doubts about, for example, equipment compatibility. When these factors are considered, together with the paucity of engineering resources that has already been mentioned, it can be seen that this approach is far from attractive.
A much more convenient and more certain solution is simply to source all of the key infrastructure elements from a single reputable and experienced supplier. Working with a single supplier brings big benefits. Engineering requirements are minimised, as all of the items of equipment will have been designed to work together. Compatibility problems are eliminated and efficient interoperation is guaranteed. This means implementation will be fast and commissioning time minimal.
If problems should be encountered, there can be no divided responsibility, which is all too often a source of delays and cost overruns. Further, a company offering a complete infrastructure solution can see the “big picture” and will often be able to put forward beneficial ideas that would not be feasible for companies offering only partial infrastructure solutions. In fact, the best suppliers may provide comprehensive design and project management services, in effect delivering a turnkey package at far below conventional turnkey costs.
This is all very well, of course, for projects like new-build data centres where the engineering team is able to specify every aspect of the power chain. But what about refurbishment and expansion projects, where some of the equipment is already in place? In such cases it is still a good strategy to choose a supplier that can provide as many as possible of the elements needed for the project in hand. These suppliers will recognise the need to accommodate infrastructure elements from third party sources and will be able to provide sound advice on how this can best be achieved.
Having established the benefits of working with a single supplier for IT infrastructure elements, let’s move on to consider the individual elements and, in particular, to examine some of the latest technologies that specifiers should look for in order to be sure that they enjoy the highest possible levels of efficiency, reliability, flexibility and value for money.
Low-voltage switchgear is the foundation of the IT power chain, but it often receives scant consideration from IT professionals. This is unfortunate, as major switchgear faults can result in protracted loss of the utility supply. Fault-tolerant switchgear with provision – in large installations at least – for switching between alternative utility supplies is, therefore, highly desirable.
Withdrawable functional elements are a major benefit as they enable maintenance and upgrades to be carried out without service interruptions. In addition, the switchgear should be designed to allow fast, easy and cost-effective expansion to accommodate future changes in requirements.
The next key element in the IT power chain is the uninterruptible power supply (UPS) system. The latest UPSs are well adapted for use in cloud-based and virtualised environments, and are supported by powerful power management software that integrates easily with all modern operating systems and virtualisation packages.
Big gains in UPS operating efficiency have been made in recent years, with models that incorporate Energy Saver System (ESS) technology delivering efficiencies as high as 99% when power quality from the utility supply is good. This translates into large savings not only in power costs, but also in cooling costs, especially in large data centres.
It’s worth noting that ESS never compromises the protection provided by the UPS. If the utility power quality deteriorates, the UPS switches to full double-conversion mode in less than two milliseconds, a transition so fast that it is completely invisible to even the most sensitive IT equipment.
Other important benefits of the latest UPSs include easy capacity testing to verify condition of the batteries, and hot-sync paralleling that provides enhanced battery management, inherent redundancy and a scalable architecture that can readily be adapted to meet increasing power requirements.
After the UPSs, the next elements in the data centre power chain are enclosure power distribution units (ePDUs). While these may seem relatively minor components in the overall scheme of things, are well worth thinking about as they can have a major impact on power reliability and operating convenience.
ePDUs are now available, for example, that provide dependable plug retention even with inexpensive standard IEC power leads, thereby removing a surprisingly common cause of interruptions in the power supply to individual servers. Depending on the type selected, ePDUs can also offer remote switching, allowing wayward servers to be conveniently restarted without needing to physically access them, and comprehensive metering that allows the power consumed by each individual server to be accurately recorded as an aid to effective power management, billing and apportionment of costs.
Finally, it’s worth remembering the important role of racks and enclosure systems in data centre infrastructure. Even though they are not strictly part of the power chain, they can significantly influence performance and reliability. The best enclosure systems provide enhanced air management to eliminate the risk of hot spots that might lead to equipment failure, and incorporate versatile security features like rack-based electronic access control. They’re also flexible in their configuration and easy to expand. In short, they’re the perfect complement for the power chain elements we’ve discussed in this article.
Without doubt, the world of data centres is a challenging environment in which to operate and, as we’ve seen, the challenges are more likely to grow rather than to diminish. Fortunately, IT professionals working in this environment can now avail themselves of a strategy that will help them to address the challenges conveniently, comprehensively and cost-effectively.
That strategy is to source as many of the key elements of their data centre infrastructure as possible from a single supplier that has proven expertise and experience. When this is done, optimised infrastructure is guaranteed, ensuring that even the most ambitious targets for energy efficiency, reliability, resilience and future expandability are within reach.