Data Center Codes And Standards
I don’t suppose that many of you would expect the phrase “moral imperative” to appear in a serious data center related discussion but, sometimes, technological capability must be tempered by human considerations. Now don’t get me wrong, we are decades away from the days in which the disposability (literally) of human capital was viewed as a calculable expense on the part of many firms, but even today, we must sometimes step back and ask ourselves if our processes and procedures remain in synch with our ability to protect our own employees. In terms of data centers, architectural changes and the increasing concerns of downtime prevention and speed of implementation increasingly call for us to evaluate if we are successfully providing for the physical protection of our employees through our adherence to published standards.
THAT’S A LOT OF ELECTRICITY
If we attempt to simplify the data center environment into its primary components, we find that it consists of two basic elements: the raised floor and the power systems used to support it. Think about it. We typically “size” data centers in terms of their power capacity —“it’s a 3MW facility”— and what this translates into is that there is one heck of a lot of electricity coursing through these places. And, since raw electricity and the human body are often incompatible with each other, a number of work place standards have been developed by regulatory agencies such as OSHA and standards organizations like the National Fire Prevention Agency (NFPA) to promote safe working environments.
DEFINING A SAFE WORKING ENVIRONMENT
Over the past few years, new standards have been developed such as the National Electrical Code that covers areas such as higher voltages, battery installation, and even modular data centers. The NFPA through its published standards 75 (covering risk assessments, aisle containment, and fire detection/suppression) and 70E that specifically covers workplace electrical safety has also provided clear guidance for data center operators on how to operate a “safe worker” environment. The problem: too many data centers don’t abide by them.
ARC FLASH: A REAL COST OF NON-COMPLIANCE
As a brief refresher, the National Institute of Occupational Safety and Health (NIOSH) defines an arc flash as “the sudden release of electrical energy through the air when a high voltage gap exists and there is breakdown between conductors.” While this certainly doesn’t sound good, when viewed from a real-life perspective, the results can be absolutely catastrophic with the production of enormous pressure, sound, light, and heat. For those unfortunate enough to be anywhere near an arc flash can see heat reach 35,000°— four times hotter than the sun’s surface. Those same workers can also be exposed to molten shrapnel and burns. Vision and hearing loss can also be common physical results of the aftermath.
The logical question to be asked at this point is why is arc flash becoming such an outstanding issue? Rather than a single reason that can easily be identified this increasing risk of arc flash incidents within the data center are the result of a combination of factors. From an architectural perspective, the increase of phased modular data centers has increased the risk of arc flash due to the adding of equipment, post commissioning, to live (energized) backplane without a shut down. This issue is exacerbated by the high cost of downtime that increasingly has generated a desire on the part of facility’s operators to make changes or modifications without “shutting anything down.” Thus, more work is performed in a live environment than can safely be justified, with the line between troubleshooting and actual maintenance in a live environment becoming increasingly blurred.
Fortunately, preventing the potential for arc flash incidents is more a function of education and process than physical modifications to the site. For many organizations this means staying current on the arc flash assessment portion of their electrical safety procedures. These assessments should require specific guidance on what equipment can be worked on, by whom, and in what state. This level of specificity provides the foundation required by many data center managers to ensure that potentially hazardous activities are correct and that safety is the primary consideration for the work being performed as opposed to making potentially hazardous decisions in the desire to minimize downtime.
Ultimately, per NIOSH, “The organization has a responsibility in preventing arc flash injuries … Organizations have the duty to provide appropriate tools, personal protective equipment, and regular maintenance of equipment and training. A commitment to training is a commitment to safety.”
As previously discussed, the business pressures associated with the avoidance of downtime often combine with the design limitations of facilities to be the primary causes of accidents that can be harmful to the site’s operational personnel. Unless a facility is specifically developed to support concurrent maintainability, equipment must be shut down to safely perform maintenance activities. Unfortunately, many data center providers attempt to mask this requirement through what can only be described as “creative marketing.” Terms like “phased builds” or “multi-tiering” all amount to the same thing — the site has a single backplane that requires that all attached data centers or modules must be brought down to avoid having to work on energized equipment.
TIER CERTIFICATION MATTERS
Since concurrent maintainability (CM) is the only way to allow maintenance activity to be performed on a piece of equipment without having to take down the entire site, Tier III certification (that requires CM) is becoming increasingly important. It is important to note that certification in this case applies to the actual constructed facility as opposed to just its design. Many providers obtain Uptime Institute (UI) certification for the paper plans, but deliver a final product that no longer meets certification requirements. Due to these disparities between what is marketed and what is actually delivered, prospective end users should insist on receiving the actual construction certification documentation from UI.
The vast majority of existing data centers are not concurrently maintainable or standards compliant. Correcting these situations will require significant upgrades. In some instances providers may elect to take their chances by maintaining the status quo within their sites. The decision to do so will be risky indeed, presenting operators and endusers with a variety of potentially adverse consequences including: down time, safety breaches, poor publicity, increased regulatory scrutiny and fines, or much worse.
Obviously, the demands of the data center industry continue to press providers and end users to maximize uptime and deliver capabilities faster than ever before. These requirements should not be incompatible with safe operation. Standards exist for a reason and compliance should never be a matter of convenience. There is no higher responsibility for any data center provider than the safety of its employees, and that is … a real moral imperative.
Chris Crosby is a recognized visionary and leader in the data center space and the founder and CEO of Compass Datacenters. Chris has over 20 years of technology experience and over 15 years of real estate and investment experience. For the first 10 years of his career, Chris was active in international and domestic sales, sales management, and product development at Nortel Networks, a major supplier of products and services that support the internet and other public and private data, voice, and multimedia communications networks. Mr. Crosby received a B.S. degree in Computer Sciences from the University of Texas at Austin. Chris is also an active member of the Young Presidents' Organization (YPO).