Because many of today's information technology (IT) systems are designed with a distributed architecture, featuring networked servers, peripherals, and interconnection devices located in a common area or scattered throughout a facility, end-users often make an incorrect assumption about powering and protecting such critical infrastructure. If the systems themselves are spread out, many assume a similar power approach should apply. So while installing a group of UPSs in the bottom of each equipment rack may appear to be the easiest approach on the surface, a real-world analysis of total cost of ownership shows this mindset may be flawed. For most deployments involving more than a couple racks of equipment, taking a centralized power approach may not only streamline requirements but also eliminate interruptions to overall operations and save money in the long run.
As electrical contractors work on the wiring, rigging, and installation of a centralized UPS system, there is little disruption to IT and other functions beyond a one-time deployment of the new power infrastructure. Once the power work is complete, the IT staff can focus on their jobs with power essentially built into the data center. Typically, a distributed system takes about one day per rack to deploy with most of the work performed by IT personnel, who generally may not be as familiar with power issues as they are with their core systems. Because of the ongoing nature of the work, multiple systems can be down frequently over many days as different components are powered up and down.
Safe access and space requirements. As data centers moved away from mainframe and minicomputer installations, IT managers adopted distributed network architectures, which make extensive use of rack-based computers and peripherals. Because power equipment does not require the same frequency of access as other devices, IT managers should be encouraged to install power protection devices in safer, less expensive areas. In fact, power may actually be more secure when located away from heavily accessed areas.
A centralized UPS located away from traffic areas can reduce the total space required for the overall installation by eliminating the need for rear access. While centralized UPS systems may seem large, the aggregate space requirement for distributed systems may end up being even larger and occupying a considerably larger footprint. In fact, rear access requirements, which are common to most distributed systems, may possibly triple the space requirements necessary for the installation. When this space is located in prime IT areas, the real cost of an installation increases rapidly.
Interconnect cabling is another hidden cost of distributed installations, where end-users often find that they are given only a single-source for proprietary interconnect cables. Meanwhile, in a comparably sized centralized installation, skilled contractors install standard wiring in accordance with the NEC after a competitive bidding process or under a maintenance contract with the end-user.
Distributed power may not be the answer. Distributed power installations can be more expensive, on a cost-per-kilowatt basis, than a centralized installation (Fig. 1 on page 20). Although a centralized UPS, oversized to provide for future growth, may be fractionally more expensive to build and install, expansion of a distributed system may be extremely costly in terms of equipment purchase as well as downtime and disruption of power during subsequent installations. In a distributed installation, with dozens and even hundreds of individual UPS systems, end-users may end up paying for what amounts to duplicate or multiple user interfaces as well as packaging, cabling, and mechanical features that may have negligible contribution to reliability.
Because planning ahead is essential for any large installation, you should make accommodations for proper wiring, distribution, and circuit breakers — all of which are extremely expensive to add after the fact. Whether the installation involves one 200kVA UPS or 100 2kVA systems, you must size the electrical service to the area for the total current required. Clearly, it's certainly not practical to increase the overall service to the facility in 2kVA increments. In most cases, neither is adding UPS systems in a piecemeal fashion. Rightsizing centralized power from the beginning usually permits smooth installations with virtually no disruption to nearby equipment.
Higher efficiency results in lower cost. The efficiency of a power system is almost invisible to the end-user. However, the savings resulting from reduced operating costs of a high-efficiency UPS can equal the operating costs of the entire power system in three to five years (Fig. 2 above). Not only does lower efficiency boost utility bills, the wasted power shows up in the form of heat that requires added air conditioning, increasing infrastructure costs and raising utility bills.
The energy efficiency of a UPS is the difference between the amount of energy (as documented by the utility meter) that goes into the UPS versus the amount of useful energy that is available to power the connected equipment. All UPS systems lose some energy in the form of heat when current passes through internal UPS components. Centralized systems typically have a 2% to 10% efficiency advantage compared to distributed systems. The fact that UPS systems operate 24×7 means that even a small improvement in efficiency can translate into tens of thousands of dollars in savings in just a year or two.
In addition, centralized UPS systems are designed to maintain efficiency with different types of loads and varying load levels. Distributed systems often specify high efficiency when fully loaded but typically operate at a fraction of their rated capacity, where their efficiency is considerably lower.
MTBF considerations. While marketing materials may boast superior reliability, engineers are aware that the statistical reliability of a given piece of equipment can be calculated by multiplying the number of components by the mean time between failures (MTBF).
Kilowatt-for-kilowatt, one large UPS has far fewer components than a distributed configuration, and those components are far more robust than those in multiple small UPS units needed to supply the same power. Even if the smaller devices have a higher individual MTBF — which they don't — the increased number of devices may lead to an inherently less reliable system overall. Because the large UPS incorporates more robust components and a rugged system design, centralized systems can have greater fault-clearing ability — as much as 200% of rated output — which leads to inherently higher reliability. Power disturbances that may shut down lesser systems are barely noticeable in a centralized UPS.
Battery maintenance and configuration. The weakest link in any UPS system, batteries have a longer life expectancy in centralized systems. In fact, it's not uncommon to see batteries in large systems last five to seven years, as opposed to three to four years in distributed systems.
Due to different cost structures and size constraints, batteries in centralized systems are often drop-shipped directly from the battery manufacturer and tend to be of higher quality than those in smaller UPS systems. In addition, centralized battery banks are generally designed to allow regular inspection and maintenance without disruption to connected equipment. The combination of higher quality batteries and ease of preventive maintenance is an important factor in increased battery life.
In addition, the double-conversion technology typically used in centralized UPS systems minimizes battery cycles compared to line-interactive technology used in distributed installations, which cycle batteries to provide power conditioning. In any battery, fewer discharge cycles translate to longer battery life, playing a key role in reducing total cost of ownership.
Even with the best batteries, careful monitoring and proactive maintenance are critical to assuring uptime. It's easier to monitor and maintain a centralized system than dozens of separate battery installations. Since every installation contains multiple batteries, monitoring multiple battery systems can be even more challenging than monitoring multiple UPS systems.
In centralized environments, battery monitoring is simplified because more sophisticated monitoring options are available for large systems and also because the cells are typically in one place rather than tucked into every equipment rack in the facility.
Final thoughts. The fact that a single small rack-mounted UPS seems inexpensive and easy to install — with little or no assistance from a skilled electrical contractor — can lead end-users to a false sense of economy.
In reality, the reduced space requirements, robust design, industrial-strength components, greater operating efficiency, optimized battery systems, and lower cost per kVA seem to support the use of centralized configurations (Table below). When the complete installation, operation, maintenance, and upgrade costs are considered, the savings and enhanced reliability of a centralized, professionally installed UPS architecture can be considerable.
Merrill is vice president of MGE UPS Systems, Inc., Costa Mesa, Calif.
|Centralized UPS||Distributed modular UPS|
|Installation||Performed by contractors familiar with power; minimal disruption to connected IT equipment||Performed by IT staff; takes approximately one day/rack to install; IT equipment down during installation|
|Space requirements||Uses no IT rack space; can be located away from heavy traffic areas; minimal access required; semi-annual preventive maintenance; front access only||Occupies prime IT rack space; usually requires rear access and larger overall footprint when total facility usage is considered|
|Rightsizing||Lower cost per kVA, given equivalent total kVA ratings (i.e. one 200kVA UPS vs 100 2kVA UPS units)||May have lower initial cost by installing fewer systems; quickly becomes more expensive with any changes.|
|Expandability||Recommend buying UPS with anticipation for inevitable growth; relatively easy and inexpensive to add batteries and power distribution||If facility was not sized for increased service, upgrades can be especially costly and disruptive. Sharing power between racks is not practical, so it is difficult to optimize power for each rack.|
|Efficiency||94% at full load and as high as 95% with typical partial loads. A 5% increase in efficiency can save more than $3,000/yr for a 100kW load.||As little as 83% at full load and potentially even less with partial loads. Lower efficiency also produces heat that requires added HVAC.|
|MTBF considerations||Lower parts count; superior MTBF; documented reliability||More complex; systems can be more failure-prone|
|Maintenance||Easier and less time-consuming to maintain with single location; better maintenance equals increased uptime||Multiple locations add to maintenance requirements; less maintenance may increase downtime|
|Service||Factory, on-site service available||Limited factory service; depot or third-party service may be offered|
|Battery maintenance and configuration||High-quality batteries in a centralized location allow easier maintenance, yielding longer battery life and lower cost.||Higher volume of smaller batteries scattered throughout the facility may make battery maintenance complex and costly, yielding shorter battery life and higher battery replacement cost.|