Data center infrastructure management (DCIM) is an emerging form of data center management that bridges the gap between traditional facilities systems and information technology (IT) systems, thereby providing operators with a consolidation of the data gleaned from each.

Although the acronym DCIM has been part of the data center management lexicon for a few years, the sense of urgency around developing a truly comprehensive software application solution is a relatively new phenomenon.

It has been fueled by an increased focus on managing efficiencies and costs related to data centers, as the footprints of these specialized facilities have been growing to meet the explosion of and heavy reliance on new technologies — including handheld devices and tablets.


In a 2007 study, The Green Grid proposed the use of power usage effectiveness (PUE), thereby providing data center operators with a high-level benchmark pertaining to the energy efficiency of their portfolios and also allowing them to compare the results against other data centers, and therefore to determine if any energy efficiency improvements need to be implemented. Since its introduction, PUE has received broad adoption in the industry.

Establishing PUE as a common metric led to the development of display dashboards and similar graphical elements, which are ostensibly employed in order to give data center operators critical information at a glance. The DCIM market thus far has been focused on providing these visual effects — and while these graphics are certainly eye-catching, they have proven limited in terms of practical application, i.e., providing operators with the data points they need.

While PUE and these other metrics accelerated the introduction of DCIM tools, which has been a necessary step in the evolution of data center management, I am convinced that the focus now needs to shift more toward managing data vs. collecting and displaying it.


The issue at the core of the DCIM puzzle is stranded data, which is part of what I refer to as the overall data challenge. Think about this from the perspective of an operator: Historically, an operator of a data center, whether on the facilities/infrastructure side (power, cooling, etc.) or the IT side, has a whole series of specialty systems that provide a variety of data points — not only separated between infrastructure and IT, but even isolated within each category.

If an operator walks into a modern, fully functional data center today, there are a multitude of systems including the building management system (BMS), the emergency power and generator control systems, the uninterruptable power system (UPS), and the electrical power management system (EPMS). But these systems do not communicate with each other and therefore it is not possible for an operator to view the information they provide in a meaningful way, in a truly connected sense.

Furthermore, because of their proprietary nature, these systems tend to be isolated from the operator’s firmwide network, which makes it difficult to access them remotely. There are tricks that will allow an operator to do so, but those are tricks vs. enterprise appropriate solutions that grant unfettered access to information. In short, there are myriad systems that collect a significant amount of data, but the data is often stranded. Even today’s most sophisticated solutions process and display data, but they have limited historical and predictive capabilities.


Key to designing a comprehensive DCIM solution is recognizing the breadth and diversity of the available information, and then developing a hierarchical system that allows you to store and analyze individual data points. In fact, there is the potential to have multiple hierarchies of data across the various systems employed by a data center. Among the pieces of information that are being collected are relationships — and defining these relationships is critical to managing them.

Think in terms of an email exchange service: An operator is running an application of some type and the application always runs on an IT device — that is a firm relationship; the IT device always sits in some type of a footprint such as a rack; that rack always sits in a pod; that pod always sits on a floor; that floor always sits in a building; that building always sits in a state; that state always sits in a region; and that region always sits somewhere on the globe.

Once you build a structure around the hierarchy of data, and when it has been populated with the information a data center operator requires, you will then be able to develop a repeatable system.


Conventional knowledge, until recently, has determined that the deployment of a successful DCIM solution would require the implementation of specialized software, hardware, and sensors — with the promise of being able to accommodate a common, real-time monitoring and management platform for all interdependent systems across IT and infrastructure. Therefore, to date, vendors have approached DCIM as a hardware problem, offering a variety of specialized devices and appliances as solutions. But DCIM is not a hardware problem, it is a data problem.

In short, any DCIM solution attempts to further the alignment of IT and facility management disciplines by centralizing monitoring, management, and intelligent capacity planning of a data center’s critical systems. Essentially, its goal is to provide a significantly more comprehensive view of all of the resources within a data center — from mechanical, electrical, and plumbing systems that form the backbone of a facility’s infrastructure to the servers and racks that compose the heart of the IT setup.

Ultimately, a great deal of intelligence will be imposed on these structures as well as highly specialized automation capabilities to create dynamic infrastructures that can actually self-adjust or tune themselves to more closely match data center resource supply with workload demand.


Rapid growth of technology has made the data center environment a larger part of the overall bottom line for a firm. And for the folks charged with monitoring these environments, the data center operators, simply collecting data isn’t good enough anymore.

Data center operators today require a DCIM system that will provide increased visibility into their operations — not just reams of individual data points, but data points that are interconnected because individual data points are not as valuable as data points presented with context.

The relatively nascent DCIM market will continue to evolve and the most comprehensive solutions built around the requirements of data center operators surely will be the game changers.