Jeff Sturgeon is Vice President, Marketing and Solutions at Emerson Network Power, and is responsible for developing, enhancing and implementing marketing programs, channel strategies, and solutions offerings for the Liebert products division of Emerson Network Power.
Emerson Network Power
As the business drivers of today’s enterprises evolve, so too do the demands placed on data center infrastructures. IT management is challenged by an array of objectives—ensuring availability, increasing efficiency, managing increased capacity—while adhering to budgetary limitations. With data centers now firmly established as high-priority business assets, how IT departments approach physical infrastructure strategies can influence not only IT effectiveness, but also business performance.
While conventional approaches to data center infrastructure design can be based on best practices and thus effectively balance efficiency and availability, for spaces requiring greater customization, they can also be resource intensive. Integrated infrastructure architectures are emerging in a variety of configurations to address organizations’ business needs while increasing efficiency, maintaining availability and speeding deployment.
Recognizing the challenges facing data center managers and the importance of adhering to infrastructure best practices, these “integrated” infrastructures offer rack, power, cooling and management systems optimized for a variety of needs and application scenarios. Using an integrated infrastructure streamlines product selection while maintaining design flexibility and providing advantages over conventional infrastructure architectures in the areas of efficiency, cost of ownership, simplicity and control.
Analyzing Infrastructure Costs
While there is much that can be said about the benefits of an integrated infrastructure in the areas of efficiency, simplicity and control, for the purposes of this article we will focus on the all-important cost of ownership.
Consider the cost of providing 160 kW of data center capacity using both conventional and integrated designs. The conventional design consists of 27 racks in three rows across 1482 square feet of raised floor space. It uses perimeter precision cooling units to deliver air through the raised floor and utilizes a central UPS system. The integrated approach uses non-raised floor space and applies row-based power and cooling with busway power distribution and intelligent aisle containment to create a solution in which power management, cooling and airflow can be optimized.
The two approaches are illustrated in Figure 1.
Figure 1. Conventional versus integrated approach to providing 160kW of data center capacity .
In terms of square footage, the integrated approach uses 41 percent less floor space than the conventional approach. Because the integrated approach enables increased rack density it uses significantly fewer racks compared to the conventional approach. By removing power and cooling constraints that prevent racks from being fully populated, the integrated architecture eliminates an entire row allowing IT and infrastructure configurations to be completed faster, enhancing deployment. Together, these factors contribute to an 11 percent savings ($56,000) in room costs over the conventional approach. The ability to support higher density racks also contributes to the efficiency improvements detailed below.
Using a row-based managed power distribution busway system instead of a room-based PDU also makes the integrated approach more cost-effective compared to the conventional data center. This generates an additional savings of $24,000.
From an energy consumption standpoint, the use of cold-aisle containment, paired with row-based precision cooling and intelligent control technology maximize the data center’s cooling efficiency, requiring 27 percent less energy annually compared to the conventional design. Extrapolated over five years, these efficiency gains result in operating savings of $85,000. These savings, paired with the CAPEX savings in room and infrastructure equipment costs, result in a significantly reduced TCO for the integrated approach compared to a conventional data center design – a cost savings of $165,000 (Figure 2).
Determining When to Use Integrated Infrastructure Architectures
The 160 kW scenario described here represents just one example of an integrated infrastructure solution. Integrated architectures are available for almost any data center challenge from single rack and row configurations to mid-size aisle-based solutions to large room architectures. Generally speaking, larger systems are structured to meet the specific requirements of the project. For example, an integrated power infrastructure may be specified based on availability and efficiency objectives and then matched with one of several complementary cooling architectures.
There are several critical questions to consider before choosing a solution.
- What is the expected capacity? Row and aisle-based integrated solutions provide CAPEX and OPEX savings up to about 400 kW of capacity. A room-level approach is best used in data centers above 400 kW.
- Do the changes impact the entire data center or just one section? Even in large facilities there may be an opportunity to use an integrated aisle-based infrastructure in one zone or section to support higher density requirements. For example, an existing 500 kW data center seeking to add 250 kW of capacity could optimize two aisles using the approach outlined previously, enabling capacity to be added within the existing square footage.
- Can you leverage existing space, cooling and power infrastructure? Often, the most cost-effective approach to infrastructure upgrades is to leverage existing systems where possible. Rather than replacing a room power system to support a capacity upgrade, it may make more sense to supplement the existing system with a row-based UPS, assuming the current system is meeting availability and efficiency objectives. This is also true for cooling where existing perimeter cooling systems can provide capacity and room balancing support for row- or rack-based systems that support higher rack density.
Intelligent, integrated infrastructures can offer IT managers simple solutions with significant savings in energy usage, space utilization and CAPEX compared to conventional data center designs. These savings result from standardizing systems components for interoperability, pre-configuration of the solution designs and strict observance of design best practices intended to maximize efficiency without compromising the availability of data center operations. While these solutions can streamline product selection and configuration, they don’t negate the need for an experienced data center specialist during the design stage. Integrated infrastructures give data center designers a new tool to optimize new and existing data centers in the most efficient way possible.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.