Converged infrastructure. Hyperconvergence. Centralized resource management, consolidated systems, increased resource utilization rates, and reduced costs.

Sounds like one of the hottest trends in IT, but in this case, it isn't. 

It's the new mantra of data center facility management as the Internet of Things (IoT) takes hold.

Rather than IT servers, data storage devices, and networking equipment, converged infrastructure is the bones and muscles of any mission critical facility — HVAC, security and safety, and critical power management systems (CPMS).

IoT is facilitating data center infrastructure management (DCIM) and can relieve a host of facility managers' “pain points.” Optimizing data center performance for efficient use of equipment and floor space, improving power reliability and efficiency (PUE and DCIE), ensuring operational continuity, managing the increasing complexity, multiplicity, and sizes of facilities — from motherships to mini's, and improving reporting and compliance are examples.

The overall magnitude of change can be mind boggling. Gartner1 projects there will be about 25 billion IoT-connected devices by the end of this decade.

Just last year, Cisco CEO John Chambers said the IoT will grow to be a $19 trillion market during the next several years. Earlier this year, IBM committed to investing $3 billion on an IoT division over four years that will concentrate on the enterprise. Big Blue also is constructing a cloud-based open platform that will enable clients and partners to build IoT solutions.

As IoT continues growing, data center infrastructures will need to keep pace in terms of compute capacity, sophistication, and required resources. Otherwise, data center owners and users could see the $19 trillion opportunity predicted by Cisco's Chambers evaporate. From a critical power perspective, keeping pace is essential to ensure efficient consumption for IT equipment and HVAC, and to minimize downtime to meet SLAs.

Those embracing the IoT know about operational issues sooner, make decisions faster, and take action more insightfully. The inescapable truth is that DCIM as a professional discipline needs and wants the change. In fact, it may be unavoidable.



The “things” part of the IoT is components, devices, and products, such as power meters and basketballs (really), that become “smart” when integrated with a “technology stack.” Multi-layer, micro-infrastructure stacks comprise sensors, microprocessors, compute capability, data storage, batteries, wireless network connectivity, and even embedded operating systems.

 The devices have local intelligence and compatible, two-way communication pathways, and, ideally, streamlined network topology protocols that eliminate repetitive wrapping and unwrapping of data.

While smart technology has been available for a while, innovation has lowered its cost, making it economically viable for widespread application, thus fueling IoT growth.

One example that's practically passé is home lighting that can be controlled from a smartphone. Another are the basketballs mentioned earlier. With a built-in technology stack, they collect data on shooting arc, dribble intensity and speed, shot release speed, imparted backspin, and other factors. A smartphone with the necessary app displays the results.

Ralph Lauren, Inc. says the company will debut a polo tech shirt this year that "tracks and streams real-time biometric data directly to your smartphone or tablet." Amazon has taken IoT a step further. It sells field-retrofittable “dash buttons” that do not have to be integrated into a product. An example is a Tide detergent button that can be stuck to a washing machine. When the consumer is low on Tide, he or she presses the button, which automatically orders more Tide, or whatever product it represents, from Amazon. Low-power microcontrollers and wireless connectivity make it possible.



Can smart DCIM components and devices be far behind? Some say they are, in fact, already here. Smart HVAC and lighting devices and components facilitate management of those systems. CPMSs for data centers are realizing IoT's potential today. Smart CPM components adjust to their environments or operating condition. Rather than run to failure, they call for maintenance.

Even though IoT-enabled CPMSs are here, it doesn't mean they have universal acceptance. Facilities decision makers responding to a national survey² about CPMS monitoring and control shed light on the capabilities they have and those they want.

More than two-thirds of respondents either have, or would like to have, monitoring capability from their CPMS. More than half either have, or need, control and reporting capabilities from their CPMS.

Almost half of those who have control and reporting capabilities also have some sort of integrated system to manage it. About 45% of respondents have some type of power quality monitoring and analytics.

No doubt, there's room for improvement.



That will materialize as smart products penetrate more facets of critical power management systems, as well as HVAC, lighting, safety and security, and building management systems. That dynamic already is creating interconnected facility management systems called “clusters.” The clusters, comprising hundreds or thousands of sensors, can be designed for a single building, multi-building campus, or geographically dispersed facilities.

Each cluster features detailed monitoring, measurement, and control capabilities, and each feeds overview and status information to an overarching building management system (BMS). The BMS orchestrates policy decisions using the aggregated data. Such capabilities make facility networks look like IT networks. 

It wasn't always that way.

Not that long ago, Ethernet serial networks with web access were cutting edge. Only HVAC, fire, and security had connectivity with building management systems, but interaction among products wasn't widespread. The range of products and their capabilities varied widely. Some may have had simple status annunciation, while others may have included monitoring capability. Most provided only data. On-the-fly analysis was left to operators.



But even IoT-enabled clusters have room to improve. To truly optimize IoT capabilities, clusters are already evolving into “ecosystems.” It's a facility infrastructure's version of IT hyperconvergence. While convergence represents bundling a variety of components into a cluster, hyperconvergence includes a management interface for components designed to work together.

Perhaps the best-known example of an ecosystem is Apple products — the iPhone, iPad, iPod, accessories, and apps. Designed to provide an exceptional customer experience, the Apple ecosystem makes consumers unlikely to switch to an Android ecosystem. If you don't think this is important, consider Blackberry. The company didn't really develop its own ecosystem and lost customers in droves, while Apple continues dominating its market, even with fierce competition.

Ecosystems are shaping up in the data center arena, too, as some critical power management system manufacturers have made inroads on building their own. Mission critical facility managers can evaluate such systems as a total solution for managing a data center cluster.

Whether they're clusters or ecosystems, one thing they have in common is generating enormous volumes of data. Big Data. In fact, the exploding number of sensing devices that share data will continue adding volume. As devices with sensing and actuation capabilities become practically ubiquitous, global adoption of Internet Protocol Version 6 (IPv6) will be essential. Managing the volume and the variety of data that almost always have different structures and meet various standards will be challenging.



Not only will facility managers be seeing more of this data than ever before, they will be expected to manage it. That has important consequences because data can be a facility manager's best friend or worst enemy. The key is to be able to interpret it.

Interpreting data correctly and quickly is what imparts value to it. It's an integrated, three-step process: monitor, predict, and improve. It starts with giving operators accurate, real-time, quantifiable data, cycle by cycle, in milliseconds. Real-time data helps operators better understand product performance. But, they must be able to assimilate it easily, which requires dynamic visualization, that is, “translating” numerical data into graphs, charts, and even pictograms.

Amassing data over time provides analytical opportunities to identify performance characteristics of a variety of devices and components and the relationships between a building and its IT systems. Monitoring a DCIM's power infrastructure, for example, can precisely determine power usage effectiveness (PUE), cooling system energy efficiency, and overall power quality.

"Power quality analytics can be used for trending and predicting growth," said Junnaid Malik, an electrical engineer at Cosentini Associates. "You may want to know where you are experiencing current level voltage distortion. You may be a co-location facility and want to plan for growth by adding servers."



When out-of-parameter performance “behaviors” occur, operators have the necessary information to diagnose the issues and take corrective action. The real payoff, however, can be the insight gained from this experience to actually predict performance issues and decide on preventive measures by changing operational parameters, or servicing the device or component. Predicting and preventing helps make a facility's power infrastructure more reliable and efficient.

For critical power management systems, the monitor, predict, and improve process facilitates reducing energy consumption, projecting capacity requirements, streamlining maintenance, resolving operational issues, and meeting reporting requirements.

Data that has value is data that requires protection. From a facility operations perspective, its continuity needs to be assured by a self-healing network that avoids disruptions or overcomes disruptions instantaneously. From a hacking perspective, its security needs to be assured by multiple levels of encryption — application layer, native database object, network, and point to point. Programmable cryptography, digital certificates, and time stamping also should be part of a data protection program.


Bhavesh Patel is vice president of global marketing for Emerson Network Power's ASCO Power Technologies business.



  1. Gartner, "The Potential Size and Diversity of the Internet of Things Mask Immediate Opportunities for IT Leaders."
  2. National Power Monitoring & Control Survey of 15,000 facility management personnel sponsored by ASCO Power
  3. Emerson Network Power, Emerson Network Power Helps Top 10 Investment Bank Operate a Power Chain 900 Miles Away with ASCO® PowerQuest