In the early years of data centers, when computing technology was costly and temperamental, protecting hardware with critical power and cooling capabilities was the top priority. Later, during the early growth years driven by the internet boom, hardware protection was a given, and data center managers were tasked with making sure availability and uptime SLAs were always met. Now, customers take availability and uptime as givens, and data center managers find themselves in a third era of data center management: optimizing energy use amid unprecedented resource proliferation and demand for a growing cadre of computing and communication services.

Rising energy costs combined with the increasing size, sophistication, and energy density of data centers present new challenges for managers under pressure to continually improve performance and reduce operational expenses. In this environment, data center managers must have access to accurate and timely information about their assets, resource use, and systems status—from granular infrastructure to high-level overviews—while also meeting operational demands in an energy-constrained environment. This level of complexity requires a new approach: data center infrastructure management (DCIM) software.

As described by Kevin Brown and Dennis Bouley in “Classification of Data Center Management Software Tools,” an APC by Schneider Electric white paper, few data center managers lack a formal system for classifying software management tools, a reflection that many data centers are managed with loosely-aligned tools for IT and facility management. DCIM, on the other hand, integrates data collection and systems control into a single point of access.

DCIM encompasses management of IT systems (computing, storage, and networking assets) and streamlining capabilities such as inventory and asset management, while providing data center managers with entirely new capabilities including change and workflow tracking, simulation of future scenarios, dependency analysis, and impact analysis of pending upgrades or facility modifications. DCIM can also integrate with software controlling the data center infrastructure, such as power, cooling, and physical security.

“Data center teams seek to satisfy growing demand for services, and at the same time optimize energy use and efficiency,” says Brown, vice president of the IT business, Schneider Electric. “In this environment, DCIM can provide a significant competitive advantage by arming managers with insight and actionable information about their assets and operations.”

DEMAND DRIVERS

The complexity of managing assets results from changes in data center architecture within new high-density data centers, driven in large part from consolidation, colocation, and virtualization. Often these changes take place amid continuing financial pressures, so managers also must often build on the existing infrastructure. Software is essential to managing this transition, and both trends are major drivers for the adoption of DCIM.

At the same time, legislation and standards within the industry are encouraging and mandating efficiency. The European Union’s Code of Conduct provides guidelines on power usage efficiency (PUE), and the U.S. EPA Energy Star for Data Centers provides guidelines on energy monitoring and efficiency. Many data center management teams now consider energy efficiency a key marketing differentiator in winning new customers who care about their environmental profiles.

Consider the scenario of upgrading a facility designed for 100 watts per square foot with a modern rack of 8 to 12 kilowatt power density, which corresponds to 200 to 240 watts per square foot—more than double the original electrical density, requiring a similar increase in cooling capacity. In this environment, DCIM reveals energy savings by integrating control of data center IT equipment with the power supply and the physical facility to help the data center team manage real-time data loads, power consumption, and, when combined with building management software, even facility security, with a streamlined set of dashboards.

OPERATIONS

DCIM enables new approaches to common data center scenarios, both in the traditional approach and in the updated process with DCIM support. Many of these scenarios are explored in “How Data Center Infrastructure Management Software Improves Planning and Cuts Operational Costs,” an APC by Schneider Electric white paper by Torben Karup Nielsen and Bouley.

The lack of physical infrastructure software can create a host of operational issues, such as:

• Lacking accurate insight into cooling demand, a mid-size data center owner oversized cooling capacity for years to maintain safety, unintentionally adding considerable and unnecessary energy costs.

• At another facility containing both low-density racks (with no redundancy) and high-density racks (with 2N redundancy), managers inadvertently installed a low-density server in a high-density rack—resulting in nearly 20 times more electrical costs.

• In a third facility, a data center operator installed nine new servers in a nearly empty rack, only to realize the next day the UPS feeding the new servers had switched to bypass mode—the overnight load had overloaded the UPS and placed hundreds of servers at risk.

 

In all these examples, DCIM software would have provided insight into existing assets, reducing human error and improving operations.

Maintenance provides another example of challenges with incomplete and siloed information. In many facilities, a wait-and-see approach dominates because of the complexity associated with maintaining a dynamic and heterogeneous environment. However, this approach can be a high-stakes gamble that puts business continuity at risk.

In another familiar situation, a monitoring system detects fan loss on a computer room air conditioner (CRAC) unit. In many facilities, the staff reacts only after the failure occurs, keeping an eye on the room temperature as a new fan is installed—creating a high risk for downtime. A DCIM-empowered system can create a work order as soon as the CRAC unit failure is detected and also compensate for the loss by speeding up nearby fans. The system can then identify servers at risk and move at-risk applications to other servers, restoring the applications to the original servers only after repairs are complete. By tracking assets, power supply, and systems dependency, DCIM software helps data center managers identify vulnerabilities and at-risk hardware in advance of adverse events—and reduces the need for significant human intervention to protect mission-critical applications when things do go wrong.

DCIM provides new capabilities in tracking IT room equipment as well. A traditional approach relies on manually logging equipment installations and audits. DCIM software can streamline this workflow, linking handheld barcode scanners on the data center floor to implementing work orders and identifying equipment. Using a wireless network, the software automatically details device and asset attributes, enabling keyword searches for equipment by name, model, and type, or exporting information into spreadsheets for status reports.

“In all these operational scenarios, a system integrated by DCIM software replaces human tribal knowledge that matured in an era of smaller, simpler facilities,” says Nielsen, principal software researcher at APC by Schneider Electric. “Interactions between assets and systems are increasingly complex in modern data centers, making DCIM necessary for managers who seek efficient and reliable operations.”

PLANNING AND CRISIS PREPARATION

Many data center managers find themselves challenged to meet new operational objectives on a tight budget. DCIM can help teams manage change and prepare managers to react to crises unique to high-uptime virtualized environments.

“The traditional planning model based on human analysis creates high risks for unforeseen consequences in today’s dense and dynamic data center,” says Nielsen. “Automated data analysis and modeling capabilities help managers understand the cascade effects of new hardware or system demands—such as how installing a new rack of servers will affect cooling capacity given existing hardware and power supplies. By integrating real-time monitoring data into planning tools, managers using DCIM can also identify ‘stranded capacity’ by coordinating assets more efficiently, in effect boosting available resources without purchasing new hardware.”

Similarly, the traditional crisis simulation model—answering “what if” questions based on manual calculations and tribal knowledge—is insufficient to plan for challenges in large facilities. DCIM provides scenario modeling for disasters, such as power or cooling failures, with scientific calculation and full analysis of systems dependency. This full view of computing resources, cooling, power, and other support infrastructure improves resiliency when (not if) the unexpected arises.

BEYOND MANAGING THE IT FLOOR

DCIM provides data center management teams with greater situational awareness, most notably in the integration of traditional data center IT assets with the power supply and the larger facility to provide insight, resiliency and control.

Within Schneider Electric’s Technology Center (SETC), the first step of integration is achieved via the Continuum building management system (BMS), providing an extremely detailed view of data center information used for calculating PUE (see figure 1). This is leveraged into a dashboard created for the C-suite, which uses the information to drive efficiency initiatives. These data may also be used for finetuning the data center’s PUE without incurring any net gain in electricity usage—in turn, keeping the balance between energy use and PUE.

The facility’s staff extends these capabilities beyond the SETC’s data center to every department in the facility, monitoring peak usage as well as tracking energy usage by department and electrical and mechanical capacity.

Accounting for the limits and characteristics of power is an integral component of modern data centers. With the rise of dense and dynamic data centers using virtualization to achieve higher asset utilization, facility power has become a bottleneck for many data centers. This is especially true with retrofits and upgrades to buildings, which may have been originally designed to handle just a fraction of the power density of new racks employing virtualization.

Additionally, the SETC utilizes Schneider Electric’s StruxureWare Central and Operations software (see figures 2 and 3) for a granular look of power to the rack power distribution level, as well as to determine limits and characteristics of the power and cooling available. This allows the management staff at the SETC to better utilize existing assets and improve planning for future expansion.

A holistic approach using DCIM integrates power monitoring to generate insight into the entire electrical distribution network, from utility feeds and transformers to the PDU and rack level. Within the SETC, Schneider Electric’s PowerLogic ION Enterprise software is used for all power monitoring within the data center—from utility to power distribution.

Often, this comprehensive approach entails integrating DCIM with the larger building control system, referred to as the data center facilities management (DCFM) system. With information from outside the data center, DCIM software can support notification and coordinate response to electrical network problems, power maintenance (planned and unplanned), and power quality and reliability analysis. Power capacity planning is crucial for facility expansion or retrofit projects and to accurately project gains from energy efficiency projects.

Data center managers can also explore integrating DCIM with DCFM software to gain insight and control over heat, ventilation, and air conditioning (HVAC), fire systems, and water, steam, and gas supplies. Such integration of data center control and monitoring with facility management software ensures proper allocation and responsiveness of all energy sources, not just electrical systems (see figure 4).

“Systems integration provides data center managers with more control, and also more visibility in the form of situational awareness,” says Brown. “An added benefit of systems integration is the ability to create energy dashboards that simply and compellingly document progress on efficiency initiatives, and also help managers identify areas for improvement.”

FACILITY AND REMOTE INTEGRATION

Properly configured, DCIM can contribute to security of the data center by feeding information from data center operations into the larger DCFM system. With DCFM software gathering information from the data center and facility security subsystems, managers can access simple dashboards providing a holistic overview from the IT room and other locations on-site or remotely. DCFM software acts upon user-set thresholds to trigger an alarm, log an event, or control physical devices—coordinated actions relying on integrating the data center with facility power, and environmental and security systems. And beyond providing visibility, this integration can empower management of the whole facility from the IT room.

In addition, DCIM offers new capabilities for remote monitoring and access. Armed with data from the DCIM system, the management team can maintain facility control through mobile clients, phone and tablet applications remotely. Unlike alerts from the previous generation of technology, DCIM provides off-site managers a complete overview of the IT space, network, power and cooling parameters.

BENEFITS OF DCIM STRATEGY

DCIM fundamentally upgrades the management of data centers to an integrated system facilitating cross-functional organization. Integration of the IT white space with power supply and the larger facility enables active energy management and managers to significantly reduce energy usage and costs, improving PUE through better asset allocation, load shifting, and operational improvements—and document their gains in clear, data-rich reports.

DCIM also improves operational efficiency, providing insight that reduces potential for common problems such as unnecessary overcooling or accidental overloading of critical systems, and it helps data center managers plan for future needs and meet challenges with a coordinated approach.

A convergence of trends is driving DCIM adoption, and as data center management teams consider how to tackle evolving challenges of efficient and reliable operations while also planning for the future, DCIM provides a powerful tool to extend throughout the entire facility for unprecedented insight and control. 

 

Reprints of this articleare available by contacting Jill DeVries at devriesj@bnpmedia.com or at 248-244-1726.