Issue link: https://cp.revolio.com/i/394834
18 | THE DATA CENTER JOURNAL www.datacenterjournal.com a lthough the importance of availability has in no way diminished, it has been joined by two new metrics—efficiency and flexibility (or scalability)—as IT managers increasingly struggle to keep pace with growing capacity needs while working under the constraints of tightened budgets and energy-efficiency initiatives. is article evaluates the costs of continuing to operate a circa-2005 infrastructure while implementing new high-efficiency technologies. It will explore how new systems can deliver 15% to 25% savings in opex, plus improved perfor- mance and a clearer path for growth. Is It worth rePlaCIng legaCy equIPMent? e first step is to assess the situation from a total cost of ownership (TCO) point of view. Begin by reviewing your current electrical infrastructure equipment for its operating efficiency at the levels where it is currently operating. Using that informa- tion you can determine the annual operat- ing expense and then compare it with potential savings gained from modernizing your data center. ere are a number of tools available that will let you look at your present equipment and compare it with these new, more modern approaches to justify your savings. Once you understand your ROI, you will be enabled to make a good business decision regarding equip- ment replacement. During the high IT growth cycles of the mid 1990s through mid-2000s, many data center managers budgeted for a 15-year replacement cycle. Data center in- frastructure technology remained relatively unchanged during that time period with no significant advancements in higher- efficiency topologies. Real advances in this area started in the 2010 to 2011 time frame through mainstream adoption of things like high-power, transformer-free uninter- ruptible power supply (UPS) technology. When looking at the overall power system infrastructures, there are two pieces to consider: the UPS system and the power distribution units (PDUs). In a data center electrical system, the UPS feeds power through the PDUs to other areas of the facility. Both of these critical components have experienced technology and efficiency advancements in the past three to five years. Up until 2010–2011, UPS technology used for enterprise-class data centers had a typical peak operating efficiency of 91% to 92%. Efficiency in this sense is a measure of how much energy can be delivered to the IT equipment as a percentage of how much energy is drawn from the utility. e other eight or nine percent was lost to heat as a byproduct of the conversion process. erefore, data center users were not only wasting energy in terms of turning that energy into heat and dispersing it into the room, but also increasing the cost to run a cooling system to push the heat out of the space. Currently, data center managers are operating the data center at partial loads owing to low infrastructure utilization rates, resulting in less efficiency. InCreasIng utIlIzatIon rate In this same time period, customers were also actively deploying more redun- dancy. Today, 2N systems are being used instead of having a single-UPS (1N) system feeding a data center, which contributes to low utilization rates. e idea behind this design premise is that a data center would never load the UPS system with more than 50% of its rating. is arrangement would leave enough capacity so that if either system failed, the IT equipment connected to the UPS would fail over to the alternate UPS and assume no more than 100% of the load. In this type of operating model, when using legacy equipment the system is not running at peak efficiency. e UPS that you thought was 92% efficient is very possibly running more in the 83–86% efficiency range. Customers will differ depending on their environment. But what were thought to be 8% energy losses contributing to the utility bill are actually 14–17% losses at the UPS alone. In the 1990s and 2000s the prior- ity for data center operators tended to be availability at any cost, as energy costs were much less of a concern. Aer the economic downturn, companies began scrutinizing their operating costs a bit more. When executives noticed that their energy costs, particularly for their data centers, were extremely high, they began looking for ways to be more efficient in order to reduce operating expenses. For some Fortune 500 companies, a 10% savings on their energy bill could result in hundreds of thousands of dollars of bottom-line savings. Because of this shi in thinking, data center managers are considering invest- ment in resilient architecture, as long as the operating costs don't outweigh the benefits. ey are challenged to look at new ways to confirm their equipment in order to improve both efficiency and utilization. Newer architectures deployed today enable higher utilization rates. Instead of having a one-to-one redundancy ratio with one backup UPS for one good UPS—which is effectively a 50% utilization rate—many customers are now using systems that may have four or more primary UPS units with only one backup UPS. ese types of architectures are referred to as reserve or catcher-bus arrangements. As UPS products have become increasingly more reliable, the data center manager can reasonably expect that no more than one UPS will fail, so there is no need to have a one-to-one ratio in your backup system. With a four-to-one architecture shi, data center managers can experience an 80% utilization rate for their UPS capacity. IMProveD oPeratIng effICIenCIes Today's modern, transformer-free UPSs provide operating efficiencies of up to 97% at full load in traditional double- conversion mode. At partial load efficiency, where the UPS is running at 40% to 60% load, the efficiencies have gone from 85% in the past to 92–95% today. at spread is significant. At partial loads, we've experi- enced as much as a 10-point improvement. At full load, we've gained four to five points of efficiency over what was the industry norm just three to four years ago. at can add up to sizable savings. For every point of efficiency improvement, a customer using a megawatt of UPS capacity can save almost $10,000 per year on utility costs for data centers located in high-cost areas. For the one-megawatt data center owner, that