Monitoring aids in data center power utilization
In the data center business, asset utilization is a key component to success. Per outlet monitoring is the latest trend for progressive thinking owner/operators and this is helping them with strategic decisions. If you don’t know exactly what the cost of running an application is, you don’t know how much to charge customers.
The rise in energy prices over the last two decades has transformed the once inconsequential cost component of doing business into a critical part of data center planning and execution. An in-depth understanding of power availability and consumption patterns within the data center is necessary for truly effective capacity planning and cost allocation.
Energy Metrics develops data center energy and operations optimization software and specialized facility and IT integration services. One of their software integration applications is the inclusion of smart power strip technology in data center installations, providing more granular monitoring and enabling the operators to gather real-time power consumption data.
Infrastructure monitoring and control
Energy Metrics uses an enterprise level data acquisition and historian to monitor measurements from HVAC, temperature control, chilled water systems, and electrical distribution systems. The system’s integrated data acquisition provides robust, real-time business intelligence for the operation of the data center. This system is a highly-scalable and secure infrastructure for the management of operations data and events.
With the rising cost of energy and the competitive nature of data centers, the ultimate goal is better asset management. “Typically the facility operations guy gets the electrical bill, and if he can’t accurately determine whom to charge, it remains his expense,” explains Ken Morikawa, VP, Technology at Energy Metrics. “If they can track their asset utilization, then they can accurately charge business units for their energy usage.” In order to do this, smart power strips are used for per-outlet monitoring across the entire system. The rack smart power distribution units (PDU) capture values such as volts, current, or watts, providing more granular monitoring of energy down to the cabinet unit level.
The data aggregated from the PDUs is used to map application energy consumption and to benchmark cost models. This allows facilities operators to know the energy usage required to run an application and then allocate these energy costs accordingly.
Real-time monitoring and notification
One of the challenges to successful implementation of a comprehensive smart power distribution monitoring system is the real-time monitoring of large amounts of power consumption data. The accuracy of each data point must be verified, outages must be recognized, and power usage levels must be compared to known capacities.
In order to accomplish this, Energy Metrics relies on EXELE TopView alarm management and notification software. The TopView software offers a direct and efficient connection to the system allowing large scale monitoring. “The benefit to using TopView is not just the ability to run multiple instances for scalability” said Ken Morikawa, “also the ease of deployment and its ability to handle a large number of data points.”
When you start monitoring different classes and categories of equipment, row and zone categories, and each outlet per cabinet, the numbers become very significant. In one installation the system monitors over 750,000 points, with 250,000 in the cabinets alone. In a typical system, every monitoring point has the potential to be in different states. “We are monitoring to the specific outlet or breaker,” explains Morikawa, “so it has to be easy to map the alarm management application to the data.”
It is essential that operations know at all times where they are compared to capacity. TopView is used to keep stakeholders informed at all times. Alarms are categorized based on severity. For example, at 80% of maximum load capacity a “warning” notification may be sent, whereas at 90% a “critical” condition message may be used. A tiered email distribution system can be setup within TopView so that different contact lists can be used depending on the time of day/week and the severity level of the alarm conditions. Unacknowledged alarms follow a configurable escalation plan. This helps in management of capacity and timely initiation of corrective action.
The first installation using this smart monitoring design using EXELE TopView software was piloted in 2011 in a small facility in Utah. As a result of the success of that system, the same customer has since had Energy Metrics install a system at their flagship data center monitoring a growing percentage of over three quarters of a million points. The client was able to reduce risk of outage by half and is able to perform more accurate capacity planning gaining productivity
Even with the increased visibility to alarm conditions and energy consumption data that this system offers, there is still untapped potential. TopView stores and provides access to the history of all alarm events through internal and external tools. At present, this reporting capability that TopView offers represents an opportunity for future improvement of data center operations. Customers have yet to embrace the trending and historical data reporting available for things like root cause analysis. “The nice thing about the potential for operational improvements using this reporting is that it already exists in the current system and doesn’t really require any further infrastructure investment,” said Morikawa.
The needs for managing asset utilization, tracking capacity, and monitoring per unit energy consumption are not going away and if anything are becoming increasingly important. The upfront investment in a more intelligent infrastructure monitoring and control system offers many tactical and strategic benefits. Improved system reliability, maximum load capacity alerting, and the ability to accurately allocate energy expenses to the appropriate cost centers or customers can offer ongoing long term paybacks for data center operators wishing to succeed in a competitive and changing environment.