Friday, November 6, 2015

You Have to Measure Before You Can Improve

Power In The Data Center

You Have to Measure Before You Can Improve


By Ed Higgins

Power Monitoring is the first step to savings in your Data Center.

As global competition intensifies, companies are increasingly turning to technology to help turn mountains of data into a competitive edge. With soaring energy prices and the need for round-the-clock data center services, enterprises must find ways to increase energy efficiency and reduce costs. In addition, escalating power consumption by large data centers and the population in general means additional power is not always available to expand computing services. Although power is becoming the most significant cost in running a data center, most data center managers lack the tools to accurately measure power consumption. All of these factors, along with a growing concern for environmental stewardship, are forcing the need for better power-monitoring technologies.

The cost of power is increasing. Power is now the single largest operating cost in the data center. The impending Carbon tax are forcing companies to truly understand their energy use patterns to reduce power usage due to increased costs associated with energy consumption. With today’s limited IT budgets, any energy savings means more money for revenue generating activities that can help bolster the bottom line.

Power failures are expensive and detrimental to business. The high cost of data center downtime due to power failure is another threat that plagues data center managers. Any downtime of the equipment in data centers supporting today’s global companies and organizations can mean millions of dollars in lost revenue, as well as withering customer confidence. Data loss or corruption resulting from power-related issues is equally damaging to a company’s revenue and reputation.

Why It’s Important to Monitor Power

You need to measure it before you can fix it. Analysts continue to rank energy efficiency as the number one concern of data center owners and operators. The truth is, however, you simply can’t improve something, especially energy efficiency, if you’re not measuring it. Energy efficiency projects often pay for themselves in energy savings, but if you don’t know how much energy you’re using and how much it costs, it is very difficult to justify new technologies and best practices or to assess the savings of those new methods. Without a baseline and then continued measurements, it is impossible to determine where to optimize, to evaluate the results of the optimizations, or to show the improvements to management, government agencies, or customers. In addition, you need to be able to identify energy consumption peaks and lows and determine how they relate to operations and key internal and external events (such as marketing campaigns, accounting cycles, and changing weather patterns) to enable you to adequately plan for these events.

A number of organizations, including The Green Grid and the Uptime Institute, are working to develop standards to help companies become more energy efficient. The Green Grid’s Power Usage Effectiveness (PUE) metric is becoming a standard for data center energy efficiency, but PUE cannot be reasonably determined if energy consumption cannot be measured. Measuring at the device plug (after all of the power conversion, switching, and conditioning is performed) is the best way to calculate PUE. Finally, measuring at the device plug is sometimes the only way to accurately measure power usage in a data center—particularly if the data center shares power with other areas in the building.

By measuring power usage you can:

• Identify potential cost savings and set goals
• Identify current power costs and set a baseline
• Implement efficiency improvement projects
• continuously measure to determine success
• Accurately bill departments and tenants
• Balance 3 phase power systems

Data center managers need to understand it to work with it. In most data center today, data center management and facilities management are still handled by two different departments, which means data center managers operate without fully understanding the ramifications of infrastructure changes.
Another reason to monitor power is to avoid costly downtime and loss of data. Systems consuming an inordinate amount of energy might be signalling a performance problem. On the other hand, inadequate power can cause stability problems. The ability to monitor power usage provides yet another tool to help data center staff actively solves potential problems, thereby possibly saving millions of dollars in losses.

The Benefits of Power Monitoring and Management

Increase profitability with lowered energy and operating costs. Even a small drop in energy consumption can deliver substantial cost savings over time.

Ensure accurate chargeback. Collocation providers charge tenants for energy usage. Monitoring power provides accurate data on usage, making it easier to compute charges. Clients are more likely to accept these charges if they are provided with accurate statements. A full accounting of energy usage can also help departments understand how effectively they are using the compute resources they purchase. Power usage data is also enabling service providers to implement a different pricing model determined by such categories as power factor. Tenants with legacy equipment having a bad power factor are charged at a higher rate for excess power usage. This will increase the Collocation provider’s green credentials.

When you understand power usage in the data center, you can also begin to intelligently balance phases on your 3 phase power system to optimize energy use and reduce costs. There are a number of benefits that make efficient load phase balancing a worthwhile objective. One such instance would be increased feeder capacity. The loading on a feeder section is synonymous with the most heavily loaded phase and, in the case of significant imbalance, feeder capacity is used inefficiently. Balancing between phases tends to equalize the phase loading by reducing the largest phase peak while increasing the load on the other phases. This equates to releasing feeder capacity that can be used for future load increases without reinforcing feeder conductors.

Additionally, phase balancing reduces feeder losses because any phase peak reduction affects the losses for the phases as the square of the current magnitude. A feeder section with 1-ohm resistance that has phase currents of 50A/100A/150A will have 35kW in losses. When balanced at 100A/100A/100A, the loss reduces down to 30kW. The same effect is even more evident in the reduction of reactive power losses because the X/R ratio of most feeder sections is greater than 1.

Phase balancing also improves the voltage on a feeder by equalizing the voltage drops in each phase along the feeder. This released feeder capacity provides more reserve loading capacity for emergency loading conditions. It is realistic to assume that the benefits in improved use of feeder capacity and improved voltage quality are of more significance than the value of loss reduction except when loading is already high.

Typically, balancing is accomplished by selecting the phase of the supply for each load so that the total load is distributed as evenly as possible between the phases for each section of feeder.

In summary, the only way to achieve power savings in the data center is to first actively measure current and power at as many granular points as is reasonably possible. From there, you will begin to see the areas of low-hanging fruit (typically the fruits represent 20%) for which you can implement strategies for cost reduction, whether this be consolidation, virtualization, elimination, or advances tactics such as integral power-capping vehicles which instruct systems to "slow-down" when right conditions are met.   Only the best DCIM solutions can provide the real-time monitoring capability to give you this insight.

I hope you enjoyed this article, and hope it was helpful.

Until next time,
Ed



Tuesday, September 8, 2015

Data Center Infrastructure Management... Facilities Management? IT Management? Business Management?

DCIM: What does this mean to you?  Part 1

By Ed Higgins

This is first in a three-part series on DCIM and relevance to facilities, IT, and the business.

First a few questions.

When the words "data center infrastructure management" or "dcim" comes up, what's the first thought that crosses your mind?

A) It's about Facility Management (e.g. the building, the power, the cooling)?
B) It's about IT Management (e.g. the servers, the switches, the storage, the virtual machines, the applications)?
C) It's about Business Management (the folks that pay the electric bill, pay for new data centers, acquire complimentary businesses)?
D) All of the above?

Now scroll down
 |
 |
 |
 |
 |
 |
 |
 |
V

If you guessed "D", then right you are!  Take a bow, for it is the responsibility of Facilities Management, IT Management, and the Business (three entities with different priorities) to work collaboratively as stakeholders in DCIM, whereby the data center operations must have adequate power and cooling, IT assets are run reliably to effectively align, and must execute to the requirements and cost objectives of the business.  The business involvement is critical, for it establishes the priority basis for which information and ranking relative to the requirements most and least important to the business.

Facility Perspective.

Being able to know where power, cooling, and floorspace capacities exist or where it is required, or where the hotspots or deficiencies are, or how much capacity will be needed in the future are considered to be the "Hot Topics" for any facilities experts who confront these topics every day.

When a power or cooling engineer can readily "see" where power or cooling capacity are abundant, or lacking, or about to change, then they can begin to align strategically and work tactically to anticipate the business's requirements.  Furthermore, when they can see further down the road, like 5 years further as an example, then they are truly in alignment to the direction of the business.

Imagine how a facility engineer feels when his environment is creeping towards 60 - 70% capacity and he learns that the business will bring in an acquired business of roughly half his company's existing size.    He'll likely say, "we need a bigger facility".  In some cases, this is the only way. But in more cases, he may have plenty of capacity stranded by inefficient and antiquated consumers of power, cooling and floorspace.

Wouldn't it be nice if the Facility engineer could anticipate these requirements, evaluate where stranded capacity exists, collaborate with IT to evaluate the impact of replacing costly consumers of capacity to free up all that which is stranded?  Well they can.


IT Perspective.

We have to deploy how many application servers into the already crammed facility?  Really?

Stay tuned for Part 2 regarding the IT Perspective, followed by Part 3 relating to the Business Perspective.

I hope you enjoyed this article, and I hope it was helpful.

Until next time,

Ed