DCS: The APC name is synonymous with data centre power/cooling – what are the ups and downs of being the market leader? 

NR: The greatest benefit of being a market leader is having a huge set of demanding customers who are always sharing their problems with you; this is where our best new ideas and opportunities come from.  One painful difficulty with being a market leader offering world-wide products is that we must take a long time in our product development processes to ensure and test for compatibility with a wide array of global performance, safety, and environmental standards, which some smaller competitors can avoid.

DCS: APC obviously likes to sell power and cooling equipment, so how do you view the current industry focus on virtualization and improving the efficiency of computing equipment?  Will customers need less power and cooling equipment? Is this a threat to your business?  

NR: In this regard it is very important to distinguish power from energy.  Many of the technologies that improve compute efficiency, such as server power management, reduce the overall energy used by reducing the energy used by IT equipment during frequent periods of low compute requirements.  These approaches do not reduce the peak power used by the IT system, so the power and cooling systems must still be sized for the peak load.  Dramatic savings in energy do not necessarily correspond to a reduction in the size of the power and cooling equipment used, so improved IT efficiency does not adversely affect our power and cooling equipment business, although it does create new technical challenges.

DCS: In future IT systems that are virtualized or have extensive use of power management, does the power and cooling system need to behave differently? 

NR: Definitely.  In the past, IT equipment was generally a constant load, with the measured power use in a data centre varying only a few percent over the course of a day.  In a data centre dominated by servers that are virtualized and/or have aggressive power management features enabled, we can expect dramatic variations in power use over the course of a day.  Not only will the total power vary by 50% or more, but power may “move around” the data centre between cabinets.  This will require that power and cooling devices be coordinated to ensure loads can be met, and capable of dynamic response to optimize the efficiency of the data centre.  

DCS: Is it right to link power and cooling so closely whenever talking about the data centre environment? 

NR: Absolutely.  Power systems process energy flowing into a data centre and Cooling systems process that same energy being removed.  These systems must always be in balance.  Imbalance results in overheating if cooling is not keeping up with power,  or in cooling overcapacity with severe energy efficiency penalty.   There is a huge benefit to linking the systems from a control and management perspective, especially in virtualized environments where power can move around within a data centre. Linkage between power and cooling controls can allow prediction of where cabinets might overheat, and can also be used to optimize cooling system efficiency.

DCS: And do your customers understand the nuances of the topic, or are you still needing to educate some? 

NR: What we find is that the very largest customers who build and operate many multi-megawatt data centres have become educated and quite comfortable with the changes such as close-coupled cooling and modular designs.  Very small customers are also accepting of the newer high density, high efficiency approaches.  However, there is unfortunately a large group of mid size customers who continue to build traditional raised-floor, low-density, low efficiency data centres because this is what they, their architects, or their construction companies are familiar with.  We spend a lot of time helping these customers understand how they can better optimize their results.

DCS: Your parent company,  Schneider Electric, has announced the EcoStruXure power and cooling architecture for buildings, which sounds a lot like APC’s InfraStruXure concept.  What is this and how will it benefit data centre customers? 

NR: For 10 years APC has been evolving a modular, integrated power, cooling, and security management solution for data centres, called InfraStruXure,  which allows users to obtain integrated, out-of-the-box functionality like capacity management without a lot of custom programming.  We made a lot of progress but the benefits were mostly limited to the computer room itself. 

We realized that to achieve our objectives we were going to need to have seamless integration with building systems such as building automation systems, and facility-level power systems. The Schneider EcoStruXure system is a holistic approach to building design and management that uses web services architecture and standard reference designs to make it much easier to design, specify, and implement high performance buildings of various types, including data centres and office buildings with internal data centres.    This is way beyond what APC alone was capable of doing, but Schneider Electric brings all the capabilities necessary to accomplish this, which will be a major advancement.

DCS: These sound like sophisticated products; are these developments for large data centres only? 

NR: Quite the opposite.   In fact our stated goal is to allow operators of small data centres to achieve the same performance and capabilities of the operators of the largest data centres.  The only way this is possible is to dramatically simplify the specification and implementation process for data centre infrastructure, and to provide out-of-the-box software functionality without the need for complicated configuration, programming, or software maintenance.  We will offer standard, tested reference designs from small server rooms all the way up to multi-megawatt data centres.

DCS: Are there any areas where you feel that you need to make up ground on competitors in terms of technology? 

NR: I feel we have industry-leading technology in all aspects of data centre infrastructure devices, such as power and cooling systems.  But our entire industry still has a long way to go in simplifying the entire data centre life cycle experience for our customers.  For example, in the areas of design optimization and of providing plug-and-play modular power and cooling capacity.

DCS: Do you see a shift in new data centre construction from data centre ownership to co-location or hosting?  

NR: Compute providers are getting very good at their business and have learned how to make standardized, modular, high performance data centres, and the advent of cloud computing architectures makes it easier to transfer compute capacity to an outsourced provider.  Nevertheless, we still find most customers would prefer to operate their own data centres, for reasons of security, control, and because they still are looking to IT as a means to achieve competitive advantage. 

We need to make it simple for these customers to design, specify, operate data centres, and reduce the cycle time associated with deploying data centre capacity.  And we believe that a properly engineered small customer-owned data centre can be just as efficient and environmentally friendly as a large hosting facility.  If we at APC and Schneider Electric do our job and provide standardized high performance, easy-to-use designs, then our customers indicate they will continue to own and operate their own data centres.  This is a top priority for us.

DCS: There is a general concern that data centre operators may find that they are unwanted in locations such as cities and other locations where power and water systems are stressed.  Do you see this as a growing problem? 

NR: This is a real problem. We can dramatically reduce water consumption in data centres, but energy use remains a challenge.  Certainly any data centres in such environments will be pressured to be extremely energy efficient, and should expect high energy costs.  The majority of data centre operators don’t see local generation as a practical alternative at this time.  In general, data centres operate with very high energy use without creating many jobs, a combination that community planners will not find attractive.  These factors will push data centres closer to sources of energy and away from population centres.

DCS: Various forecasts suggest that the power density of data centres will continue to increase.  What do you see happening? Is there a practical power density limit? 

NR: The power density of data centres will continue to increase because a more compact data centre consumes fewer resources such as copper, steel, and concrete, and is fundamentally more energy efficient because air flow paths are shorter and better defined.  Although chip technology improvements are reducing the power consumption per core, the industry will use these improvements to pack ever-more computing power into the same physical space. The natural practical limit for air cooled data centres using rack mounted equipment is in the range of 20kW per rack or 7,000 W per square meter (or even higher in specially engineered environments such as containers).  The average data centre today operates at only 1,000 W per square meter, which is well below the practical limit, so we can expect a continual increase in power density in the next few years.

DCS: Over the past ten years the industry the data centre infrastructure industry has gone from talking about reliability, then density, and now efficiency.  Have we solved these problems?

NR: I really think we have technically solved the density and efficiency challenges for data centre infrastructure, it is only an issue of adoption and implementation.  With regard to the reliability issue, marginal improvements continue, but the major advances in overall system reliability will be obtained through improvements in IT system resilience and not via facility infrastructure.

DCS: What is next? 

NR: From an environmental perspective, we can expect a trend to decrease or eliminate water consumption in data centres, and pressure to manage and reduce the total life cycle consumption of natural resources and the embedded carbon in both facility infrastructure and compute devices. 

We will also see a relentless drive toward standardized modular designs for physical infrastructure of data centres.  Custom engineered one-of-a-kind data centres will be viewed as hopelessly outdated, unmaintainable curiosities.

ShareThis

Tags: Power & Cooling, Green Tech

Read more White Papers»

More DCS TV»

More Audio»

Read more Web Exclusives»

Read more Magazine Articles»

Read more Supplements »

Keep up to date with the latest industry products, services and technologies from the world's leading IT companies.

• Virtualization Success Depends on Network Automation by Infoblox

• The Infoblox VMware vCenter Orchestrator Plug-in by Infoblox

• Automated Storage Tiering on Infortrend’s ESVA Solution by Infortrend

• Why Cloud Backup: Top 10 Reasons by Autonomy

• How to Measure the ROI of Cloud Data Protection by Autonomy

Latest IT jobs from leading companies.

Click here for full listings»