By Martin Olsen
Published in the June 2006 issue of Today’s Facility Manager
Blade servers, computer room air conditioning units (CRACUs), hot and cold aisle floor plan configurations, Moore’s law—these terms are becoming increasingly prevalent in the facilities management vernacular. Never before have the roads of facilities management (FM) and information technology (IT) professionals converged as much as they do today.
Smaller servers with increased processing power are coupled with a wide variety of business applications. Everything from e-mail, Web hosting, and enterprise resource planning applications to access control, fire detection systems, Voice over Internet Protocol, and workforce management systems can be housed and processed on the same servers, which drives the demand for more efficient, flexible, and secure data center solutions.
Consolidation of functions is resulting in increasing heat and, in turn, skyrocketing electrical costs for today’s businesses. Consequently, the time is right for FM and IT professionals to offer holistic solutions to the issues surrounding reliable business application availability and uptime in an efficient and cost-effective manner.
A New Paradigm
Given the fact that IT professionals are now demanding a minimum of 8kW to 10kW of cooling capacity or more per enclosure within the data center, traditional data center design should be renamed from “room to rack” to “rack to room.” Large corporations have filled their raised floor space to capacity with rows of rack mounted servers, storage, and network equipment capable of supporting ever increasing processing power and speeding information to customers and employees worldwide.
With each data center potentially operating at 8kW to 10kW, there are significant cost implications. Therefore, FM and IT professionals must jointly address these implications within the context of the company’s bottom line.
The Importance Of FM Intervention
Despite the necessity of the department’s involvement in data center planning, FM is often excluded until late in the process (if included at all). However, FM professionals must become stakeholders early in the process. They must approach their IT counterparts and become familiar with the different enclosure (or rack) platforms available.
Today’s enclosures go well beyond simple cabinets that solely house servers. They have transformed into sophisticated solutions which feature integrated technologies that more efficiently store, cool, power, manage, and secure mission-critical hardware. Essentially, they are mini data centers that reside within the company’s physical data center.
As facility professionals educate themselves regarding rack offerings, they should be wary of anything that appears to be pre-packaged or proprietary. Although a “one size fits all” approach may insinuate standardization, it could seriously limit options.
A truly scalable enclosure offers an open architecture or design that adapts effectively with any third party networking and server hardware, regardless of the manufacturer. Only a few manufacturers of enclosures and data center infrastructure solutions can honestly claim this type of third party compatibility.
And while a multi-vendor compatible enclosure may initially be more expensive, it could offer standardization capable of driving out mistakes on the data center floor, particularly since the platform will accommodate whatever equipment is being purchased.
The prominent question of how to cool the rising heat loads in today’s data centers is the subject of much discussion. Some solutions may make significant promises in terms of thermal management, but others may change the world of FM and IT professionals forever.
The notion of thermal management can be very daunting, since it is an intangible concept to some. Cold air distribution (as opposed to heat removal) is not the only answer to thermal management issues, especially when the following factors are considered:
- Obstructions under the raised floor (this can create non-uniform airflow through floor tiles);
- Re-circulation of hot air from the rear portion of the enclosure to the front of the server air intakes;
- Lack of a hot aisle/cold aisle strategy where enclosures are all facing the same direction (this can cause an effect where units “drink” heat exhaust from each other);
- Stratification across the front face of the enclosure (this can lead to failures in top mounted network and server equipment); and
- Exhausted heat from the enclosure that never returns to CRAC units.
The Next Generation Of Data Centers
Research indicates that today’s data center planners are preparing for their next generation of upgrades, redesigns, or rebuilds to address the increasing power and thermal management demands of high density technologies.
All data center planning should start with the enclosure and build out from there. Deploying a scalable rack platform that stores, cools, powers, manages, and secures a number of different server brands and models will make all the difference when moves, adds, or changes (MAC) come around.
And since temperature is such an important issue in the data center, enclosure platforms should be capable of adjusting from low (up to 3.5kW) to high (more than 5kW) per enclosure without having to change the primary platform. Redeploying servers and cabling can run from $50 to $75 per component, adding up to thousands of dollars for every enclosure change out.
Recovery rates are important to any cost conscious manager. The more equipment that can be recovered during MACs, the better the initial investment.
By making careful decisions, FM and IT can free up data center personnel to concentrate on core competencies. This can result in a winning scenario for all.
Olsen is director of data center marketing for Wright Line LLC of Worcester, MA; e-mail your questions about this article to [email protected].