Issue link: https://cp.revolio.com/i/833286
12 | THE DATA CENTER JOURNAL www.datacenterjournal.com Although industry organizations provide overall guidelines and stan- dards for maintaining data centers, managers must take their approach a step further and consider their facil- ity's particular design, equipment and operational requirements when devel- oping a maintenance plan. It's crucial to remember that industry standards outline the bare minimum of care and that all managers should evaluate and implement a care plan that goes well beyond those core requirements. e process for developing these site-specific standards begins with as- sessing three factors: 1. Risk: First and foremost, with any data center, it's critical to under- stand the business's risk appetite. e standards a data center operator puts into place directly correlate to the fa- cility's risk tolerance. When assessing risk, several elements require consid- eration, such as the type of industry the business is in (which aligns with how critical the data center is to the overall business), the technology be- ing used, the data center's location and how closely data center operations are tied to business operations (i.e., colocation and cloud providers). 2. Inventory: Once the risk has been determined, the next step is to inventory all data center assets to understand what needs protection. When conducting the inventory as- sessment, it's important to note details such as the age, make and model of the equipment, as they will be es- sential to establishing maintenance requirements. 3. Manufacturer recommenda- tions: Similar to carmakers, data cen- ter equipment manufacturers provide a recommended maintenance sched- ule for processes, such as changing the generator oil or uninterruptible power supply (UPS) batteries, to avoid high-cost repairs and extend the life of the device. Data center managers shouldn't take these recommendations as the end-all, be-all for servicing their equipment but should factor them into their comprehensive planning process. Only aer these three elements are clearly outlined and reviewed can a data center manager start to prop- erly map out the standards that will set the foundation for their mainte- nance plan. data ceNter StaNdard deSigN aNd implemeNtatioN At a high level, establishing stan- dards begins with understanding the core requirements for each piece of equipment in the data center. On the whole, equipment should perform the same basic functions regardless of the manufacturer. For instance, all UPSs should do two main things: provide emergency power to a load when the input power source or main power fails, and hold that load until the gen- erator and/or utility power kicks in. Once the equipment fundamentals are defined, standards can then be built around those core requirements. e next level requires applying the policy standards to each site in the data center portfolio. Unfortunately, this task tends to be where the "wheels fall off." As such, all site managers should map out how the high-level standards apply to their particular fa- cility to help them develop and tailor specific methods of procedure. Here's an example of how the process would go for a data center UPS system: e company standard for UPS maintenance regulates that all units must be serviced twice a year. One of the company's data centers is a 2N facility that has redundancy built into the UPS systems, so the site man- ager would need to establish specific equipment-by-site requirements to service one of the UPS systems while performing a switching activity that transfers load to the redundant sys- tem. e site manager would create a scripted work effort that maps out the switching and transfer sequence as well as how they would perform maintenance on unit A without shut- ting off the critical load to unit B. Once they have determined how to seamlessly transfer the load from A to B without interruption, they have cre- ated a procedure that's repeatable both in frequency and technique, ultimately reducing the risk of UPS failure due to human error. One important point to remem- ber is that there is no one-size-fits- all approach to creating data center standards. In some cases, identical facilities can have varying standards depending on their risk profile. For example, a data center located in a hurricane-prone area might test emergency and backup power systems bimonthly during storm season (June to November), whereas an identi- cal facility located in a region with less adverse weather conditions may only test their equipment every other month during the same timeframe. Similarly, two vehicles of the exact same make, model and year can have different maintenance plans depend- ing on the number of miles driven per year or how safely the car is driven. In both instances, what's needed is a set of unique standards and procedures that account for how the equipment is expected to perform in the context of how it's being used. Establishing and implementing standards is a laborious process and can't be done overnight. It's an im- portant investment, however, that can ensure the data center is maintained consistently and performs as expected. Whether it be cars or data centers, the key to longevity and health lies in good planning and upkeep. n about the author: Joe Reele is Vice President, Datacenter Solution Architects, Schneider Electric