Absolute Power
Typical power receptacles found on any wall may look like a secure and guaranteed source of power, but the reality is much different. When plugged into that receptacle, power-sensitive devices could fail an average of 14 times in a single year, resulting in an astonishing cost to many businesses. These power events (as termed in a politically correct manner by utilities) are typically not the result of a blackout, a popular perception, but rather from less conspicuous issues such as sags in the voltage level. Ensuring continuity of just about any modern operation goes hand-in-hand with guaranteeing power quality and guarding against the plethora of power-related bugs, which cumulatively cost businesses billions of dollars each year.
The facts are that more than 80 percent of power events are sags in the line voltage (where the line voltage drops by more than 15 percent of its rated voltage). Most interruptions last less than a second and are hardly noticeable to the human eye, but for a computer, server or other sensitive device, this is enough of an interruption to knock the device offline. Surges, spikes, brownouts and blackouts (the rarest but most severe form of power interruptions) accompany sags in the legions of power quality issues. Figure 1 illustrates the most common power quality problems and their effects.
In the case of ever-so-common sags, the culprit can usually be found within a facility. Most often they are caused from devices that share the utility line, consuming big surges of power when turned on. Equipment like laser printers, photocopiers, motors and even computers may consume upwards of six times their rated current when first switched on. It is these brief surges in demand that cause the line voltage to drop below critical levels. Computers, servers and IT equipment are usually outfitted with power supplies that have some ability to ride through sub-cycle (i.e. <8 ms) outages, but nonetheless they may be knocked offline many times a year when raw utility input is their only power source.
The cause of most power-related issues is no fault of the utility, but rather from within your facility or from nearby influences. Receiving its fair share of media coverage over the past few years, electrical utilities have been faced with a host of issues rising from deregulation. For customers, in many areas of the country, this may be painfully apparent with the dramatic price increases, brownouts and even rolling blackouts. Political statements aside, one certainty looming amidst all the talk about power is that America’s power grid is deteriorating to that of a third world quality.
Part of the problem arises from the lack of incentive for utilities to invest in the power grid, since they no longer are guaranteed a return under the current deregulated environment. Under a regulated and monopolistic environment, utilities could structure pricing in a fashion that would guarantee a return on investment in infrastructure. Adding to the issue is the fact that these lines are getting saturated carrying all of the power they can during peak periods such as hot summer days. The end result is that the energy grid is no longer as dependable as it once was, at a time when it is more critical than ever to have a reliable source of power – a costly proposition to a business environment in which 35 percent of all devices are “power sensitive.”
Unfortunately, when IT equipment is shutdown ungracefully (i.e. pulling the plug) critical data is usually lost, transactions are interrupted and manual intervention is often required to restart the device. In the case of a large data center, this can equate to an eight hour plus exercise that may result in a huge volume of lost transactions and revenue.
For today’s businesses, the cost of an interruption may have sizeable consequences. Average losses are greater than $1 million for each hour of downtime. Industries dependent on automated systems, such as energy and manufacturing, average nearly $3 million per hour in downtime losses, and less automation-dependent industries, such as healthcare and hospitality, fall in the $330,000 to $650,000 range for hourly losses, according to an October 2000 study by Meta Group. Figure 2shows the estimated costs of power-related issues.
Office Network : Even a brief lockup on a network server equates to a significant loss in productivity. Assuming 50 people rely on the server access, minimum recovery time from a power glitch is usually 15 to 30 minutes. This easily equates to $1,000 to $10,000 of lost productivity with each shutdown depending on the value of the productivity. Considering that a power protection solution for such an environment can start at less than $1,000, the ROI is practically guaranteed.
Financial Data Center: Data centers are prime examples of the extreme financial consequences of power-related problems. The advent of B2B transactions has meant that downtime will have a direct financial correlation to the business in terms of lost opportunity.
Approximately 35 percent (and growing) of electronic devices are deemed power sensitive, meaning that they need a continuous and uninterrupted source of power. The only way to protect devices from the widest range of power quality problems is to use an Uninterruptible Power Supply (UPS). UPSs are available in all varieties imaginable from inexpensive desktop units, delivering basic power protection, to multi-module systems used to protect large financial and telecommunication data centers. The following primer will show you how to select the right UPS for your application. The net result will be a fluid operating environment immune to the perils of power interruptions….and ultimately a very respectable ROI.
Selecting the Right Power Protection Solution for Your Business
Browsing through the myriad of UPS offerings, choosing the right one can seem like a daunting task, as many appear to do similar things. There are some huge differences that you need to be aware of to make an educated decision and stay protected. The best way to view the decision is to first look at the criticality of your equipment in terms of how much will it cost you should the service be interrupted. The cost of interruptions can be assessed in one or more of the terms shown below:
Customer Service: Customers unable to access Web sites, call for service due to a PBX outage or to complete a sales transaction are prime examples of lost revenue and customer good-will.
Lost Productivity: Not being able to carry out tasks that rely on electronic/IT-based business systems will directly correlate to employee downtime and/or loss of revenue generating hours. Examples range from as simple as an e-mail server going down to a bank of CAD stations locking-up in an entire engineering division.
Lost Transactions: Not being able to execute revenue generating transactions and permanently losing the opportunity. This extends to B2B transactions, point-of-sale (POS) transactions, Web-based ordering and banking and financial market transactions.
Interruption of Critical Processes: Many manufacturing processes such as pharmaceuticals and semiconductor manufacturing rely on having a continuous process. Interruptions can cause catastrophic damage to entire batches of product.
Safety Liability: Many systems can have safety repercussions when not operating properly leading to potential liability issues.
It’s important to note that UPSs are available in a range of topologies that usually equate to their cost. Should your business system be a fairly simple application such as protecting departmental desktop PCs or distributed POS terminals, you can select a solution that uses an “off-line” topology. Off-line UPSs feed raw utility power to the protected device until such time as the power fails. The UPS then switches to battery power and typically offers five to 15 minutes of reserve battery power. Such systems can be an inexpensive source of basic power protection.
For mid-range solutions such as servers and more critical centralized computer equipment, a higher-grade “line-interactive” UPS topology is recommended. Line-interactive UPSs also supply raw utility power (typically with some filtering or surge suppression) when available, but maintain some degree of regulation on the output voltage without depleting batteries, to provide improved power quality.
Larger, critical installations such as a data center require true “on-line” topology UPS systems. On-line or double-conversion topology UPSs have the advantage of regenerating 100 percent of the power that is fed to the critical devices. As such, they can maintain perfect power quality under just about all conditions. While physically larger, on-line UPSs offer higher power ratings with more power per cubic foot of installation and can be installed in non-premium spaces outside the data center. They are typically used for centralized applications (where one UPS system feeds power to an entire facility or area). Such centralized implementations also have the benefit of being far more economical and easier to maintain than distributing many small UPSs throughout a facility.
Before requesting a UPS system(s), keep in mind there are huge variations in quality between brands. You usually get what you pay for with low-end retail brands. Many low-end off-line UPSs experience premature battery failure that will cause maintenance nightmares and often more downtime than if your equipment wasn’t protected. Off-line and line-interactive systems can transfer to battery in just four milliseconds and provide basic protection against power sags and surges. Although these systems are easily installed (typically under a desk) with simple plug-and-play operation, they can pose problems in a distributed environment where it is difficult to identify which systems require routine battery maintenance or locate a malfunctioning device when troubleshooting more significant problems. In addition, factory service for these devices is generally limited, with little or no on-site support available and few third-party options. For larger data center-type systems, it’s best to go with a UPS company that has an in-house engineering group and can work hand-in-hand with a consulting engineer. These projects always benefit from the third-party insight of an impartial consulting engineer, as there is a lot more to specifying and installing a system than meets the eye. While it may seem more costly at first glance, a properly engineered centralized system invariably provides superior protection with a lower total cost of ownership.
Most UPSs have an optional communications port that can interface with your network to monitor and manage your critical power. Furthermore, the UPS integrates with most IT networks to provide a command to the servers and IT devices to gracefully shutdown in an orderly fashion should the UPS be approaching the end of its battery backup time. When the power returns, the device will automatically reboot and maintain normal operations within minutes. This graceful shutdown will result in a system being able to boot-up within minutes versus hours, should the “plug have been pulled” abruptly.
If the nature of your system dictates no tolerance for downtime, you may have to mitigate the potential of a prolonged outage by complementing the UPS system with additional batteries, or in the case of larger systems where this may not be economical, a backup generator. Again, systems with a high degree of criticality often have parallel UPSs for redundancy with the distribution systems being configured in a fault tolerant manner. Your UPS manufacturer or consulting engineer can assist you in devising the best architecture for your application.
Looking to the Future
In the dawn of the computer age, no one imagined that the meek white box sitting on everyone’s desktop would pose the next energy crises. In the early 1990s, steel mills and large factories were among America’s largest single consumers of electricity. Looking at a 50,000 square foot data center today, it may use more energy than a small steel mill. Proliferated millions of times throughout the country, IT infrastructure and even PCs add up to an estimated 30 percent of the nation’s energy consumption. Together they are the new digital steel mills. Combined with a growing overall electrical demand, many parts of the country experience days when the electrical reserves are teetering dangerously in the red, on the brink of rolling blackouts.
There are slow and steady initiatives to address our growing electricity demand. Among the many solutions is distributed generation, whereby individual companies, or one day even individual houses, operate their own micro generating station. Technologies like natural gas-powered microturbines (tiny jet engines that produce quiet and efficient power) already enjoy a decent installed base and are slowly growing in acceptance. Other technologies include fuel cells that, like other emerging technologies, are buffered from everyday use by a hefty price tag. So for now, we are at the mercy of the utility to keep us plugged in, but prudent managers with well-designed power protection systems need not worry for now.
Alan Katz is Three Phase Product Marketing Manager at MGE UPS SYSTEMS, INC., 1660 Scenic Drive, Costa Mesa, CA 92626. For more information, call 800-523-0142 x7317 or e-mail alan.katz@mgeups.com.
