Data Center Trends: Monitoring Data Centers

Server room
Photo: Stockbyte.

By Bhavesh S. Patel
Published in the March 2014 issue of Today’s Facility Manager

For data center facility managers (fms), having pertinent and extensive information about what is happening with their on-site power and distribution system is important. Data centers rely on uninterrupted power to conduct “business as usual” even when utility power fails. Today’s critical power systems designed to provide continuity of power at a data center should normal power fail are complex. Therefore, facility management (FM) and IT departments alike would benefit from a sophisticated monitoring and control system that aims to ensure reliable operation when called for.

There are various solutions on the market that offer different levels of capabilities for gathering, monitoring, analyzing, and acting on data from building systems, critical power generation, and distribution components and systems.

Fms can evaluate which type of system could work best for their data centers, and one way to do this is looking at four types of sophisticated power monitoring and control systems. There are two legacy systems: Supervisory Control and Data Acquisition (SCADA) systems and Building Management Systems (BMS), along with two newer system types: Data Center Infrastructure Management (DCIM) and Critical Power Management System (CPMS).

These four technologies have similarities and differences worth highlighting. Though each of the four systems can improve efficiency and operational reliability and enhance worker safety, because of their differences one technology may be more suitable for a specific installation than another.

Evaluating The Best Match

BMS and DCIM systems are designed to monitor and control both utility power and critical power for an entire facility but these do not operate at the speeds required to process large volumes of data in real time. A CPMS—which does have that speed—is dedicated to monitor, control, and analyze critical power generation and distribution systems for the entire data center and have that information available to managers. Choosing a system is not exclusionary. A CPMS will work alongside a BMS, a DCIM, or a SCADA.

A CPMS for a data center has the ability to monitor and report on the condition, operation, and status of automatic transfer switches, static transfer switches, generator paralleling switchgear, circuit breakers, load banks, paralleling bus, UPS, and other equipment using the display terminal. A CPMS also encompasses technical service and support—essential to ensuring emergency power reliability and availability. It can also interact with a data center BMS, helping fms make sure these are maintained in tiptop condition. Lack of proper maintenance is the leading cause of on-site power failure to start.

Power quality analytics is another component of CPMS systems that ensures continued operation and helps with what-if analysis or in preventing potential power problems.

Monitoring the condition of data center power involves multiple types of equipment, which are not always located within the data center space. (Image: ASCO.)
Monitoring the condition of data center power involves multiple types of equipment, which are not always located within the data center space. (Image: ASCO.)

A manufacturer may offer multiple CPMS packages with various capabilities and differences in the system’s interoperability, flexibility in configuring a power monitoring and control system for virtually any on-site power system, scalability, and centralization. The availability of such options enables fms and other stakeholders to select the best matched system for their data centers.

Meanwhile, a BMS is, in essence, a reactive system that could be used proactively by data center fms responsible for management of the critical infrastructure, alarming, diagnosis, and control of non-critical systems.

That said, it is important to understand that a BMS does monitor and control both utility power and critical power for a facility, but, not having the bandwidth, it does not operate with the same high speed as a CPMS, which looks at everything it focuses on more deeply. Unlike other monitoring and control systems, a BMS enables two-way communication between fms and their employees or occupants.

In addition to being attentive to the power system at a data center, a BMS controls, monitors, optimizes, and reports on air handling and cooling systems, including lighting, fire systems, and security systems, typically representing about 70% of a facility’s energy use. Software for a BMS may be proprietary, using protocols such as Profibus or C-bus, or open architecture that integrates Internet protocols and open standards like Lon, Modbus, XML, and BacNet.

Next to consider is SCADA, which, optimally, provides alarm handling, trending, diagnostics, maintenance scheduling, logistics management, detailed schematics for a particular sensor or machine, and expert system troubleshooting guides. The current generation of SCADA systems includes computer and open system (off the shelf) architecture that acquires data from and sends commands to monitored equipment, human-machine interface (e.g., a computer monitor screen), a networked communication infrastructure, sensors and control relays, remote terminal units, and programmable logic controllers. However, a SCADA system typically does not encompass technical service and support.

Meanwhile, a DCIM solution, which may be said to give a holistic view of data center performance, can encompass many specialized capabilities that could provide benefits for data center operators. A DCIM could focus on monitoring and managing server traffic, or it can provide synchronized management of the entire data center. It is a proactive system that gathers information from many sources. Because it is so encompassing, it enables fms to make a broad range of informed decisions in diverse areas ranging from asset management to IP performance.

A DCIM, specific to the data industry, can monitor 20, 1,000, or 100,000 points of data. However, like a BMS, it does not look as deeply as a CPMS, which is designed to make decisions at the equipment or the subsystem level. And unlike a CPMS, which is more narrowly focused and can get as deep into a layer as needed, a DCIM cannot delve too deep into any one particular layer. Otherwise, it will exhaust its processing capacity, which would adversely impact other capabilities.

A DCIM manages energy, assets, availability, risk, services, the supply chain, and IT automation by acquiring data using SNMP (Simple Network Management Protocol), Modbus, or BACnet. As a system, DCIM can encompass specialized 3D visualization software, hardware, and sensors to monitor and control all IT and data center infrastructure equipment in real time. These systems automate three primary functions: data collection, infrastructure modeling, and analytical reporting.

When implemented to optimal benefit, DCIM improves uptime, enables efficient capacity planning and management, and provides business analytics.

DCIM bridges the gap between the FM and IT departments. If purchasing one, fms and their IT counterparts should seek to agree on which capabilities are must-haves and how the information is to be shared.

Patel
Patel

It is important for a data center fm to determine what information they are looking to obtain and manage before picking a monitoring and control solution. It may be best to choose one type of system if it will provide the data needed by the organization, or with multiple systems, as long as the systems can communicate with each other (if two systems) or among themselves (if three systems). Decision-making need not be a case of “either/or” but rather of one system complementing the other.

Patel is director of marketing and customer support at ASCO in Florham Park, NJ, a business of Emerson Network Power. He is an accomplished public speaker with deep expertise in power system markets.