For many years, data centers have been vilified as energy hogs. Tons of equipment, running 24x7, with the goal of zero downtime, generating copious accumulated heat to remediate, meant high energy demand and huge cooling bills. According to the Environmental Protection Agency (EPA), data centers consume as much as 20 times more energy per square foot than a typical office building, and have traditionally been the largest cost center for many organizations. It was accepted for decades that a data center should be kept near meat-locker temperatures in order to protect the function of the equipment within.
Pressure to reduce the enormous energy waste and introduce more savings has come from both within data centers as well as from environmental groups who have been critical of such energy waste. CFOs realize that sky-high energy bills are not just a foregone conclusion of doing business, and are demanding improvement. The federal government has issued a mandate to reduce spending on its computing infrastructure, targeting data center consolidation and efficiency improvements. Mounting societal pressure includes a recent report from Greenpeace outing the biggest energy consumers and questioning whether their power is sustainably sourced. With all this negativity, what is keeping data center operators from addressing this waste? A simple answer may reside in the question, “Where to start?”
For facilities managers to improve on the performance of their data centers, they must first deploy a solution that is capable of monitoring all the power — where it’s coming from, where it’s being used, and how to improve that use.
Data center infrastructure management (DCIM) provides the real-time information needed to get a handle on power spending. DCIM provides:
Monitoring of assets, workflow, capacity, and energy usage among other metrics
Alarm features to alert the data center manager when a reading is nearing a pre-established limit
Temperature tracking, so that hot and cold spots can be addressed, and ambient temperature can be raised to safe but more economical levels
DCIM’s value has also been realized within the federal government — which operates some of the largest data centers. The March 2016 Data Center Optimization Initiative (DCOI) mandates all federal data centers install DCIM software and report on strategies to address inefficient processes, optimize existing facilities, achieve cost savings, and transition to more efficient operations. This mandate’s goals are ambitious — to save the federal government a total of $1.36 billion by the end of 2018 and achieve data center optimization, cost savings and avoidance, and closing data centers.
The overall cost-savings goal is to reduce data center spending by 25%, with a facility closure goal of decommissioning 52% of federal data centers.
In addition, the federal agencies need to be able to prove these improvements. The most important metric is power usage effectiveness (PUE), which is defined as total data center energy used divided by total IT equipment energy used. The federal goal set for PUE is less than 1.5 for existing data centers, and below 1.2 for new constructions.
To achieve these objectives, old manual methods of reporting in the data center must be replaced with new automated practices. Many facilities currently use several monitoring tools that unfortunately don’t synchronize data, providing a fragmented picture of the facility, with no central view of everything and how different systems are interrelated. Gaining this “big picture” view is vital if facilities managers are to improve the optimization process with informed business decisions.
Since many data centers that have grown organically over the years tend to have all types of IT and facilities equipment, of varying ages and manufacturers, using many different protocols, the DCIM solution selected should have the capability to communicate with a heterogeneous set of assets, no matter what “language” they speak. No equipment should be “siloed” because its information is unreachable.
A good DCIM solution can help achieve all the above-discussed goals in several ways. In order to use floor space most efficiently, the safe consolidation of assets is key. To enable this, the data center manager needs to know where the assets are on the floor, where these assets are in their lifecycle, what work they are performing, and what work has been done to them. Hence the manager needs asset management, workflow information, and an audit trail of changes made.
The data center management team needs to monitor the capacity of energy in the facility (energy metering is another DCOI goal) so it can be optimized, without the loss of uptime. DCIM can also help discover power and assets that might be underutilized, so improvements can be made to use this excess capacity more efficiently. Assets can then be placed in the most efficient configurations.
Asset tracking also reveals old and energy inefficient equipment that might be candidates for retirement, as part of a healthy tech refresh cycle. If an asset costs more to maintain and power than the value of the work it performs, its retirement should be considered. If enough outdated gear is retired and replaced with efficient equipment then facilities can be operated more efficiently and staff can be more productive.
No doubt all of this work must be conducted safely since downtime is not an option. So both real-time information and the ability to test before deploying a change to the power chain are keys and the ability to perform “what if” scenario testing can keep uptime safe and secure.
Your DCIM solution should be able to answer, in a safe testing environment, questions like:
What would happen if I moved this asset?
What if this piece of equipment failed?
Where would the load go? What else might fail as a result?
Do I have the proper redundancy?
Because DCOI is a federal initiative, security is of the utmost concern. The DCIM solution selected should be certified to be secure by a neutral third party — this ensures it has been hardened and tested for application level vulnerabilities. In addition, it should provide dashboards and reports required to prove progress and monitor success levels.
Recently, a major federal department selected a DCIM solution — after reviewing many other solutions available in the market — to help achieve its mandated DCOI goals. The agency needed a DCIM solution to monitor and manage hundreds of server racks across multiple data centers. Previously, the agency had used Excel spreadsheets, Visio files, and in-house created software tools to monitor its operations, which created data silos, scalability issues, and inaccurate information. They found this system both cumbersome and inefficient, not to mention lacking a single source view of their entire operation.
The DCIM solution that was selected successfully brought all the pertinent information together in a single pane of glass, so more informed business decisions could be made. Additionally, the DCIM’s built-in connectors and workflow capability integrated easily with the existing IT service management (ITSM) platform to give IT operations up-to-the-minute information regarding data center assets and resources.
Now the agency has a sophisticated and automated monitoring system that will help discover underutilized capacity to maximize space and energy efficiency for easy consolidation and conservation. Real-time reporting and environmental monitoring allows the agency’s management to be confident in the placement of assets within limits and without fear of overload. Older, inefficient equipment has been safely retired. Agency personnel also have real-time access to reports that can help them determine their levels of compliance with the federal mandates, which allows them to proactively plan for the future of their facilities.
GREENPEACE UNDERSCORES PUE
Recently, Greenpeace published its 2017 renewable energy report, “Clicking Clean: Who is Winning the Race to Build a Green Internet,” in which they outlined the energy footprints of large data center operators, websites, and applications. Some of the biggest and most notable companies in the world are on this scorecard, for better or worse.
To give an idea of scale, Greenpeace states, “The energy footprint of the IT sector is already estimated to consume approximately 7% of global electricity. With an anticipated threefold increase in global internet traffic by 2020, the internet’s energy footprint is expected to rise further, fueled both by our individual consumption of data and by the spread of the digital age to more of the world’s population, from 3 billion to over 4 billion globally.”
One method to measure a data center’s shades of “green”ness is via PUE, which not only tells data center management how close they are to their efficiency goals, but measuring energy usage and PUE over time helps determine which conservation initiatives are effective, and which might need improvements.
As in the case of federally mandated energy consumption improvement, a DCIM solution can help any data center seeking to be more energy efficient, shrink their carbon footprint, and prove their progress toward a greener infrastructure.
In conclusion, no doubt the pressure is on the data center management community to conserve energy and be prepared to prove the progress achieved. Questions about data center energy spending are coming from inside the data center’s own organization, from the U.S. government, and from environmental activist groups.
Energy conservation is an ongoing process where facilities’ operators must constantly monitor to ensure initiatives are working and that progress is being made.
DCIM solutions can be an integral part of this process by providing data needed for successful implementation of energy efficiency and other tasks. Once a data center manager has a complete, unified picture of the facility, and knows where the energy in the data center is being expended, steps toward efficiency improvements can easily be implemented to reduce energy usage and carbon footprint.
Progress can indeed be measured to prove tangible results, and DCIM solutions can provide the intelligence needed to achieve them. Reducing costs and creating greener data center environments are additional significant benefits of gaining operational efficiency.