33.1.1 DCIM Maturity—The Technology and the User

Any new technology typically emerges after a long journey. Taking any of the recent data center examples like Virtualization or Cloud computing, it can be noted that there are several distinct periods that must be traversed before any technology is deployed in standard Production. Gartner refers to this flow as the “Hype Cycle,” a broader interpretation of which is shown in Figure 33.2. What starts as an amazing innovative idea with all the promise in the world gets tested and retested over the subsequent years with a dose of reality thrown in for good measure. Technologists and business managers alike poke and prod at new inventions to determine how it could pertain to their own use cases. Over time, some inventions vanish for various reasons, and others emerge with general adoption growth rates over time.

DCIM has been following this same curve. Referring to Figure 33.3, you can see that various organizations are currently at different stages in their ability to think about their computing needs in the future and are being challenged to self-evaluate their own IT best practices of the present. After years and years of ad hoc asset management solutions for the data center and the typical unique change processes that have abounded throughout the data center industry, we find a set of points along the maturity continuum that characterizes the industry’s current status.

33.1.2 DCIM Is Strategically Important to the Modern Data Center

DCIM is a resource and capacity planning business management solution for the data center. It enables the data center to leverage existing physical layer technologies including monitoring, capacity planning, configuration data bases, environmental sensing, etc., and it also enables the seamless integration into an enterprise’s other business management solutions used for asset management, process management, data management, HR planning, budgetary planning, SOX compliance,¹ etc. In a twenty-first century company, information is the most strategic asset and competitive differentiator. The data center structure itself is the factory floor that produces that strategic value of processing workloads or transactions. DCIM is the manager of the data center floor in which hundreds of millions of dollars of assets and billions of dollars of information flow in nearly every fortune 500 enterprises. DCIM is the business management for this critical infrastructure and will be instrumental to the dynamic self-adjusting data center of the near future.

There are at least four stakeholders emerging as having keen interest in DCIM, and each has their own sets of needs from the adoption of DCIM: (1) the IT organization, (2) the Facilities organization, (3) the Operations’ finance department, and (4) the Corporate Social Responsibility individuals.

In general, the compilation of everything “IT” is becoming viewed as a single entity that has a quantifiable value to the organization and associated cost. The IT and Facilities organizations are now being tasked with a common set of goals regarding the data center, and in response, finding themselves required to behave as a single business unit, with transparency, oversight and accountability, forecasting, and overall effectiveness, all in focus. DCIM enables that focus and has become essential for the data center community.

Andy Lawrence at The 451 Group put it most succinctly: “We believe it is difficult to achieve the more advanced levels of data center maturity, or of data center effectiveness generally, without extensive use of DCIM software. Today, the three main drivers of investment in DCIM software are economics (mainly through energy-related savings), improved availability, and improved manageability and flexibility.”

As the IT industry transformation continues, we see a handful of corporate IT goals that are greatly influenced, enabled, and supported by a well-conceived DCIM strategy, which puts DCIM squarely in the same level of importance as any of the other deployed business management applications:

1. IT transformation including the Cloud
2. Actively managing power needs and focus on Green IT
3. Asset Migrations and Capacity Planning
4. Operational Excellence and process reengineering
5. Deferring new capital expenditures and, in particular, Data Center builds
6. Audit and Regulatory and Legislation compliance.

33.1.3 Common Goals for DCIM

DCIM has become a general-purpose “efficiency” category where a wide range of stakeholders have voiced their data center management needs. Some of the most common goals for the introduction of DCIM into operational and strategic plans include the following:

1. Energy management has become the first priority for most data center managers and IT business managers alike. Their goal is to reduce operating costs across the board, with the starting point being much more proactive energy management. DCIM solutions can provide a highly granular understanding about energy usage and a wide range of other physical layer metrics and ultimately help to identify and control inefficiencies.
2. The need for highly accurate and actionable data and views about the current capacity and availability of their data centers and easy access to baselines associated with current operations.
3. Operational Best Practices are being redefined to accommodate much more streamlined operations and remediation. DCIM is being viewed as the tool best suited to identify, present, and manage the workflow associated with data center physical assets. DCIM essentially captures and enforces the processes associated with change.
4. With such a wide range of traditional tools in use, even within the same organization, there exists a very loose and disconnected source of truth about the data center assets. DCIM promises to coordinate and consolidate disparate sources into a single view of asset knowledge.
5. Resource availability and capacity planning is high on the list of needs. Better predictability for space, power, and cooling capacity means increased useful asset life spans and increased timeframes to react to future shortfalls.
6. Keen insight allows large amount of raw data to be transformed into business intelligence. Enhanced understanding of the present and future states of the data center allows for better asset utilization and increased availability.
7. Corporate Responsibility goals to assure the latest innovations in IT management are being considered, investigated, business cases are being created, and as a standard practice, major technological advances are not being missed.

33.1.4 Whose Job Is DCIM? IT or Facilities?

One of the most interesting aspects of the adoption of DCIM is the audience diversity and their individual driving factors. Traditionally, data centers were built, maintained, and utilized by different distinct organizations: (i) the Facilities organization that took care of all space, power, and cooling requirements for any piece of real estate, and (ii) the IT organization that took care of the data center physical and logical build and all of the equipment lifecycles itself.

IT and Facilities now find themselves required to work together for planning and optimization processes. Decisions about equipment and placement are now being made jointly. In many companies, both organizations now report to the CIO. DCIM enables the data center resource capacity planning to be managed over long periods of time. That said, DCIM and all of its capabilities will (in most cases) be driven by the IT organization just as it has for every other type of data center management (Systems Management, Network Management, etc.) over the years. Sure there will be benefits across all groups once DCIM is in production, but the IT organization tends to have the largest role and experience in enterprise-class software selection, so it is expected that the deployment and organizational integration of DCIM is best leveraged as an extension to existing software management frameworks driving data center logical operations already. In a few years, as the DCIM implementations have become mature, this layer will play an instrumental role in matching supply and demand, assuring just the right amount of processing resources exist at each point in time.

33.2.1 Capacity Management including Power, Cooling, and Floor Space

Power was first major DCIM trigger on everyone’s list. The rising cost of power was being seen in all aspects of life, both residential and commercial. Individuals saw the price of gasoline and electricity rise, corporations saw their huge power bills become larger. IT is typically the largest single line item in a corporation, so the abnormal rise in these highly visible costs caused a stir. The CEO and CFO leaders began asking questions about true IT costs, which the CIO and their teams were unable to answer. Power is one of the first quantifiable values that are directly associated with a successful implementation of DCIM.

33.2.2 Business Process Reengineering and Operational Efficiency

The operation of a data center is quickly transforming from individual and disconnected tactical activities with an historical primary goal of “high service levels at any cost” to a planned and predictable approach with the modified metric “service at what cost.” Essentially, the cost factor is being added to the equation and being tested at every step of the way. IT organizations are being asked not just to document and then automate their existing practices, but to actually consider their current approaches and determine if they are still valid and/or optimal. As such, a number of organizations are finding themselves with limited awareness of their existing practices, which is impeding their ability to create streamlined new approaches. As baselines are created for existing conditions, IT organizations will begin to author new optimized workflows and deploy new technologies such as DCIM to manage their assets over long periods of time. DCIM promises to be able to capture current business practices, and allow optimizations in workflow and labor-related efficiency to be realized.

33.2.3 Data Center Consolidation Projects

Data center consolidation is reality today for most corporations for various reasons: advancements in computing technology, mergers, and acquisitions. DCIM supports the commissioning and decommissioning of vast amounts of computing equipment typically found in data center consolidation projects.

33.2.4 New Capacity, New Data Centers

Many organizations are realizing that their core data center assets are either past their useful lifespan or are simply not able to support their organization’s rapidly increasing demands for processing due in part to their inefficient practices and wasted resources. The acceleration in the adoption of new business applications was never imagined to be at the current rate, so these inefficiencies have presented themselves across the board.

DCIM promises to address the quantification of current data center capacity, with a keen eye on capacity management over time. The data center itself provides computing resources; and when a large sample of time-based usage data is studied in combination with the demands associated new corporate initiatives, highly accurate data center planning is not only possible, but expected. DCIM quantifies data center capacity and allows it to be planned.

33.2.5 Data Center Cost Reductions, Enhanced Resource Efficiency

With the era of “Green computing” came the primary goal of reducing waste. Specifically focused on energy overhead, it has become quite popular to focus the majority of data center optimization projects on their ability to allow the data center to operate at a lower cost per unit of work. Green IT has been used as a catchall phrase to describe the more efficient usage of power.

33.2.6 Technology Refresh and Architectural Changes

A good number of data centers find themselves with large-scale technology refresh projects. These projects stem from the desire for higher density computing, virtualization, Virtual Desktop Infrastructure (VDI) initiatives, or mobility. Entire infrastructures are being redesigned, and when faced with this level of change, IT professionals find themselves looking for innovative ways to manage these new designs more effectively than previously practiced.

33.2.7 Environment and Sustainability Focus

There is a great deal of interest shown by most major corporations in reducing the impact of IT on the environment. Many organizations use three key metrics proposed by The Green Grid and covered elsewhere in this book to describe their efforts toward environmental friendliness: Power Utilization Effectiveness (PUE) relating to overall efficiency in the data center, Carbon Utilization Effectiveness (CUE), which refers to the Carbon footprint associated with energy consumption, and most recently, a metric associated with water, Water Utilization Effectiveness (WUE), which represents the amount of water consumed in the production of data.

33.2.8 Regulatory and Compliance, Audit and Documentation

The executive teams within major corporations across the globe have found themselves under new levels of scrutiny regarding IT. IT as the most critical corporate asset is involved in every major function company-wide. The impact of IT has become so great and pervasive that various government and regulatory agencies are striving to provide oversight to assure that data are maintained accurately and that the environmental impact of the data center is considered.

DCIM becomes a means toward this end. DCIM allows a data center to be documented as a single system, with the intricacies of its components identified and understood. The efficiency of the operation of each component can be seen and, over time, optimized.

33.2.9 The Cloud

DCIM fits everywhere! Public and Private Clouds share a common set of characteristics: Self-Service, Quick Provisioning, and Accounting. For a Public Cloud provider, scalable DCIM solutions are required to help quickly manage assets and dynamically tune supply and demand. A well-conceived DCIM solution is essential for the Public Cloud providers to understand all capacities (across IT and Facilities) and thus allow quick-turn for remediation, for provisioning, and for decommissioning. DCIM enables the data center to run as a business with all of these costs clearly quantified and optimized. DCIM allows Public Clouds to exist, to be more responsive, more accuracy in their operations, and reduce the overhead required to provide their end-user customers’ required levels of service.

Private clouds are just traditional IT infrastructures that have been operationally transformed using the principles pioneered in the Public Cloud world. DCIM solutions are proving to be one of the most significant enabling technologies for this IT infrastructure reengineering. DCIM will allow this Private Cloud transformation. Remember that DCIM is all about enabling the data center to be managed like a business: comprehensive access to all of the business metrics, costs structures, services, etc. and dynamic management of assets. A comprehensive DCIM solution is essential for the transformation of traditional IT infrastructures to a highly tuned, optimized Private Cloud.

33.3.1 Asset Lifecycle Requirements and Change Management

DCIM enables lifecycle management of the data center and all of its assets. It addresses the physical layer of the data center and includes the same change management and workflow capabilities found in the other Enterprise Resources Planning (ERP)-class business management solutions found in the typical enterprise. DCIM is not just a monitoring utility, although there is monitoring within DCIM solutions. The biggest value of DCIM is not a monitoring utility. DCIM is as an enabler to manage change with a keen eye on the cost structures associated with this change.

Over the course of these years, it is estimated that at least 25–30% of the assets contained within any given data center change each year. Technology refresh cycles due to depreciation and maintenance costs, adoption of dense computing and virtualization, new networking or storage technologies, all account for huge amounts of change.

As can be seen in Figure 33.4, changing a single server seems relatively simple, imagine multiplying that effort by a thousand or ten thousand times each month! It is staggering. DCIM is the business management platform that keeps all of these Add/Move/Change cycles in order, documents the process at each step, and identifies the tasks needed to be completed in extreme detail to reduce human errors incurred during the execution of these tasks.

33.3.2 Capacity Planning, Analytics, and Forecasting

Of specific note to the DCIM opportunity discussion is its ability to consider the data center as a system, with a very specific set of metrics and capacity over time. Data centers have physical attributes and associates limitations. Whether it is space, power, or cooling, each data center has a physical set of limitations that define the limits of a data center’s capacity. DCIM has already been shown to be the best way to look at these factors together, and then consider over time. With this ability to consider all resources over time, predictions can be made about when one or more of these critical resources will be exhausted and what cost will it take to bring new resources online.

The most successful DCIM offerings understand that visibility into the future is extremely valuable. It’s quite easy to focus on historical data and present it in various forms, but interpreting historical data and using it to trend into the future is where mature DCIM offerings shine. Worth noting is the recent IDC findings that almost one-third of all data centers are forced to delay the introduction of new business services and more than a quarter of those data centers needed spend unplanned OPEX budget to maintain a poorly defined data center structure. These unrealized opportunity costs can be huge!

The DCIM model includes a highly granular representation of the data center, which enables it to identify where resources (power, space, cooling, and connectivity) exist and where they are being used. Over the years, many data centers have lost resources due to their inability to identify their exact location. Terms such as “stranded capacity” and “vertical white space” come into discussion when these conditions occur. Essentially, the originally designed resources become fragmented and therefore cannot be effectively utilized, or in other cases the availability of one resource is not coresident with similar capacity of another resource. A great example is a data center that wishes to deploy high-density blade chassis systems in an area with plenty of power but limited cooling. That power essentially becomes “stranded.” The same types of imbalances occur across all of the data center resources. Modern DCIM solutions help by identifying when resources exist and allow balancing to recapture these resources. In some cases, this repositioning of equipment to better balance all available resources may add two or more years of useful life to existing data center structures.

33.3.3 Real-Time Data Collection

There are two major types of operational data that must be collected. The first type is the traditional “IT” devices and their virtualization components. These devices communicate most commonly using traditional networking protocols such as Simple Network Management Protocol (SNMP) or modern web-based Application Programming Interfaces (APIs) and include fairly well-defined templates that are embedded by each data collection utility which understands how to interpret the various values provided by the device itself. These devices report hundreds of data points, so the mapping of just those values needed by DCIM is critical.

The second type of device important to the DCIM solution is all of the components that form the Mechanical, Electrical and Plumbing (MEP) infrastructure. These include power and cooling devices typically found external to the data center, or those devices used to provide large volumes of power and cooling for subsequent distribution. This includes generators, battery backup UPS systems, large floor mounted PDUs, cooling chillers, and CRAC/CRAH units. These devices typically communicate with more challenging protocols such as MODbus, BACnet, LON, and in some cases older serial command lines via ASCII RS232.

In general, data center metrics useful to DCIM solutions are observed every few minutes by polling. In a few cases, there are triggered asynchronous events like doors opening, but the vast majority of this “real-time” data in a data center relates to temperature, humidity, pressure, and power, and those metrics are measured over longer periods of time, with analytics looking for trends over those same periods. Worth noting is that “real time” in the context of DCIM is not sub-second real time as the manufacturing world might define it, but instead typically deals with metrics observed over minutes or hours.

33.3.4 Integrations with Third-Party and/or Existing Management Frameworks, Web APIs

One of the requirements for DCIM solutions is their ability to connect to existing structures. Most IT and Facility organizations have deployed point management solutions over time. These solutions have formed the core of data center management for years. The strongest DCIM solutions will be those that provide connectivity to these solutions as well as a number of traditional business management applications to coordinate workflows and metrics in a meaningful fashion. These systems can provide a wealth of knowledge source, are critical to service desk and ticketing processes, and include all of the control hooks to the existing components. There are dozens of IT and Facilities systems that will be found across the many diverse corporate data centers, and DCIM vendors increasingly find their customers asking for these integrations. Integrations range from simple device access using standard protocols like SNMP or WMI to more complex web-based integrations of workflow and power-chain management. The integrations are seemingly endless and the strongest DCIM are accumulating inventories of these integration “conduits.” Figure 33.5 shows just a sampling of major systems that will ultimately be connected over time to perform the DCIM function. Prospective customers should consider these inventories of “off-the-shelf” conduits from each vendor when making their choices.

33.3.5 Discovery Services and Location Services

33.3.6 Data Import and Export

One of the important features when implementing any DCIM solution is the ability to gather and normalize existing sources of asset and connectivity data. In a complex data center, there may be hundreds of thousands of individual pieces of data that would otherwise have to be manually entered or recreated through some means. In general, the labor cost to establish this knowledge manually without using any data import will exceed the cost of the DCIM software license itself, and in some cases may actually be twice the cost of the software license. Hence, the significance of data import innovation is a critically important part of the DCIM solution.

In response, most DCIM vendors include some means to import data sources such as spreadsheets and text files. Each vendor takes a different approach to importing and includes varying degrees of intelligence during the import process.

Most mature solutions use advanced field and pattern recognition and will even handle fairly well-defined types of problem resolution during the import process to map to existing source files. These files vary in format, and the error corrections available during the importing process may include missing information lookup, sequential missing data replacement, asset field de-duplication, proper handling of structured cabling range conventions, and a general ranking of data fields based on overlapping sources.

Data import is typically a critical component of any DCIM solution at the time of deployment. It is most commonly used once, and then the previous means to track and maintain asset knowledge are abandoned in lieu of the production DCIM solution. The most effective DCIM implementations allow the DCIM suite to become single source of truth about assets, once put into production.

In a related topic, some DCIM solutions also enable the EXPORT of data to industry standard file formats such as CSV or XLS. These exports may include some or all of the DCIM database information, and tend to be large files used for ad hoc analysis to feed into other systems transitionally. DCIM solutions that include the export functionality can usually recreate the entire main database using this same file as an import.

33.3.7 Materials Catalog and Library

All DCIM solutions are designed to manipulate asset material lifecycles, their placements, and their connectivity. In the creation of the physical structure, various types of devices must be selected from this catalog and then used throughout the DCIM modeling process.

Most vendors of DCIM solutions supply material libraries with 5000 IT devices or more. It is the extent and means to enhance this library that will define the success and ease of use when attempting to articulate the current complexion of the data center faithfully.

The materials catalog includes representations of devices and includes the manufacturer’s specified parameters for each device. These parameters typically include high-resolution renderings of the front and back of the devices, power requirements, physical dimensions, weight, connectivity, etc. In the case of complex devices, the material catalog also includes options that may be installed (i.e., power supplies, and interface cards). All of these materials must be supplied by the DCIM vendor, or must be created manually by the end-user, which is a huge undertaking. Some vendors offer the ability to request these new devices to be added “on demand” (and typically within a week or two), while other DCIM vendors require the user to create these special new devices themselves. In a few cases, the DCIM vendor provides both mechanisms to enhance the materials catalog.

33.3.8 Rack Planning and Design

One of the most visual features of any DCIM solution is the ability to create faithful representations of equipment racks and their installed gear and associated connectivity. In fact, it is the visual representations of these racks elevations that typically attract some of the most enthusiastic initial interest by data center managers regarding DCIM. When considering DCIM vendors, a great deal of weight is given to the level of fidelity and absolute accuracy of the racks created by a given DCIM solution, and each offering is judged by its ability to most closely represent their real-world counterparts.

Most DCIM solutions use the aforementioned materials catalog as building blocks for rack design as shown in Figure 33.6.

33.3.9 Floor Space Planning

The floor of the data center is essentially an X–Y coordinated grid used to identify the actual location of equipment racks and other free-standing data center gears. Floor planning is a critical reference process as the floor of the data center must be designed to mimic the actual geometries in each data center.

Unfortunately, the data centers in use today are not always in simple rectangles. There are many types of construction and various obstacles which influence the placement of equipment and racks with the data center. The DCIM solution’s floor planning component must allow these geometries to be captured accurately as they influence nearly all other aspects of modeling and planning within the data center if desired. Accurate positioning for every device and rack is a core requirement to realize the maximum benefit of DCIM.

Most DCIM offerings also include the visual representation of the data center at large using the floor planning component. Shown here in Figure 33.7 is an example of where we see various top-down representations of the data center, with the floor-tile systems, racks, CRACs, and other components shown in precise detail. These top-down views are also able to present metric data and aggregations using color-coded scales. For instance, they can represent the number of available rack-units, or total power consumption in each given rack. Using these visual representations of the data center, capacity can be visualized and new projects can be created based on the actual complexion of the data center as it sits currently.

33.3.10 Reporting, a Critical Part of the DCIM Story

One of the most valued capabilities for any DCIM solution is its reporting. Reporting is the way in which the raw information is correlated and then presented in a business impactful fashion. These reporting systems may also include a library of standard data center management reports and can typically distribute any of these desired reports to specific user(s) in an automated fashion. Other DCIM vendors simply include data store definition schemas and rely on their customers to design their required reports and then use industry leading reporting packages such as Microsoft Reporting Services, Business Objects, or SAS to create these desired reports.

33.3.11 Dashboards. A Picture Is Worth a Thousand Words

Dashboards tend to a special case of reporting and can be considered the “At-a-Glance” report. Dashboards have the ability to present vast amounts of information in easy-to-read displays suitable for desktop or operations “command center” consoles. Even though dashboards could be considered by some to be a “cosmetic” attribute of the DCIM suite, they are one of the first considerations new DCIM prospects look for when selecting a DCIM solution. Remember, the amount of data available within a DCIM solution can be enormous, and the ability to distill large amounts of this raw data into meaningful information that can then be presented using easy-to-understand visual dashboard elements is key to overall success. The ability to quickly access actionable information is a major value for DCIM deployments.

The vendor’s included dashboards (Fig. 33.8) are a critical presentation of the DCIM function. With so many stakeholders in the proper operation of the data center, each will have a set of metrics that they hold themselves accountable to. Each has a set of needs which can be measured and derived from data found within the data center. Operations and Finance look at costs of equipment, depreciation, warranty, etc. Facilities professionals look for trends in power and cooling consumption.

33.4.1 The Platform’s Architecture

In this section, we’ll describe the DCIM platform as well as the instrumentation layer that supports it. As stated previously, the DCIM data model is a living 3D model of assets, and as a general rule, the more connected this model is to the wealth of real-world instrumentation available, the higher the realized value from DCIM.

These early single-user developments were not bad choices in the pre-2006 timeframe, but most of these early DCIM applications are now being augmented or entirely replaced by well-behaved modern web-based versions that are deployed upon enterprise-class and IT-maintained business servers. “The Web” and all of the advanced communications and presentation technologies have provided a huge opportunity to create complex management applications that can be easily scaled and widely accessed from anywhere on the Internet. These modern DCIM offerings typically scale by spanning across multiple server engines, with each engine serving various user, data collection, storage, and analytics/reporting functions.

33.4.2 The Platform’s Data Storage Model

A core component of the DCIM suite is a robust data storage model. Large amounts of data will be sourced from a wide variety of sources, much of it being time-series in nature, and all of it being required to be readily accessible for complex analysis and presentation needs. The data model itself must be robust enough to store data in a way that very complex analytics can be used across the data set interactively.

It is critically important for DCIM suites to have data models that are designed for interactive retrieval. Large volumes of data will be stored over time, and one of the key attributes of a strategic DCIM solution is its ability to present interpretations of vast quantities of raw data into meaningful metrics. While it may sound like a technical detail, the choice of storage approaches will directly affect the usability of the entire DCIM solution. Users will not tolerate the slow performance caused by data retrieval in a DCIM solution. Complex searches across gigabytes of data would take unacceptable amounts of time if the wrong storage technology was chosen. Imagine that every time you wanted to use an application on your smart phone, there was a 30-s delay before the first screen. You would likely NOT use the smart phone. The DCIM storage model, if chosen poorly, has the potential to have the same effect.

33.4.3 The Platform’s User Interface

Modern DCIM suites are most commonly web-based and utilize the latest web-based access methods being adopted in common business management applications. The visual presentation of DCIM is complex and varies from vendor to vendor, but each shares a common goal of allowing easy navigation across a vast field of data by multiple users across the Internet.

The Graphical User Interface (GUI) can be considered one of the key attributes of a DCIM solution, as customer adoption many times is directly related to the intuitive nature of the GUI. A great example of an intuitive GUI interface is Google’s Earth application, which allows the untrained user to start with a map of the entire planet and within seconds zoom into a view showing the house where they live.

The user interface for DCIM is critical. These applications must be highly intuitive. Large populations of any combination of IT and Facilities equipment spanning over thousands of square feet must be quickly accessible. DCIM enables the relationship between components to be clearly articulated in great detail.

33.4.4 Instrumentation: Sensing the Physical Components in Real Time

Modern data centers can provide a wealth of information about the current status of everything from the power chain and cooling status to the performance of the servers and virtualization layers. It’s all fit into a category that is leveraged by DCIM called “instrumentation.” Instrumentation is essential for a DCIM solution to be effective, and it includes a wide range of technologies and protocols, each intended to gather a specific portion of the entire infrastructure. The most mature DCIM suites expect many of these subsystems to exist and the DCIM systems themselves deal with the normalization and presentation of this instrumentation data. A point of reference regarding scale is worth noting here, as the magnitude of data gathers using various means of instrumentation can be massive. In a typical small data center with 100 racks and 1000 servers, tens of thousands of data points per MINUTE can be generated!

33.4.5 The Rack, the Most Basic Building Block of the Data Center

Data center racks themselves are the physical building block for the IT task. Typically, a physical cabinet is made of steel, and each rack is typically 6-ft tall, 2-ft wide, and slightly longer than 3-ft deep. Commonly, 42–48 devices may be housed in each rack, with larger or smaller numbers being seen in specific applications.

As standard building blocks, most DCIM offerings understand these mechanical designs and use very accurate templates for the selected rack(s). The size and shape of each rack is well understood and when coupled with floor-tile systems, allow an extremely accurate representation of the data center to be modeled. DCIM offerings use these building blocks as the basis for their high-fidelity physical topology representations, and rely on this ordered approach when depicting location and relative placements.

33.4.6 Remote Access and Power State Management

Related to DCIM has been the notion of remote access to systems. In fact, long before the DCIM marketplace emerged, the concept of managing the infrastructure was left primarily to two constituents: (1) Facilities managers who visualized power and cooling using purpose built BMSs and control panels and (2) systems administrators who used hardware and/or software tools to access their equipment remotely to power cycle or reconfigure operational settings.

The Facilities manager’s ability to manage their power and cooling infrastructures has become quite mature. Highly advanced and completely customized visualization, dashboard, and control mechanisms have been created by the large building automation vendors. These BMSs are tailored individually for each deployment and tend to be quite functional, albeit extremely rigid. BMSs and their more advanced counterparts Building Automation Systems (BASs) have high price tags, and must be defined at the time of building construction in extreme detail by the Facilities engineers who manage the building power, cooling, security, and lighting systems. Building engineers and mechanical designers work in concert to create these infrastructures, and then BMSs/BASs are tailored to reveal the inner workings and control capabilities. These systems tend to change very little over time and only when major Facilities construction changes occur do these needs get reevaluated and capabilities updated.

For the IT world, some of these remote management technologies are referred to as “power cycling,” “KVM,” or simply “console” and can be seen in Figure 33.9. Essential physical management of IT systems and data center devices was based upon the one-user-to-one-device approach using one or more of these technologies. This brute force approach to IT device management was directed at a single server, switch, or other type of IT system as a stand-alone management entity. The servers and other devices had no notion of placement and relative location or required resources, and basic metrics for energy consumption and temperature usually did not exist. Remote management technologies can be considered some of the most basic and primitive means of device management used for the past dozen years. The need for this type of remote access has been greatly reduced by hardware and software maturity found in enterprise-class equipment and today is most commonly found where a specific mission-critical device (and single point of failure) is deployed for specific functions. For these applications, power cycling and associated system reboot is the most common use of these remote management technologies.

Although not strictly required for a successful implementation of DCIM, these suites can usually take advantage of remote access technologies already deployed. These DCIM systems can allow traditional system administrators to share the user interface found within DCIM suites, and navigate to the system where remote operator access is desired, addressing configuration or reboot requirements.

33.5.1 Selecting the DCIM Vendor

Look for a DCIM vendor with a vision that matches your own. Obviously, every data center strategy is unique across the industry, but there will be well-formed thoughts about capacity planning, operational excellence, energy management, disaster recovery, etc. that must be discussed prior to choosing a DCIM vendor.

Take into account how long the vendor’s solution has been available and how many installations each vendor has. Obviously, more is better as it supports and defends the vendor’s approach to DCIM and will ultimately help steer your choice. Each vendor’s installed base will have provided a priceless resource of users requirements from similarly situated users who have walked down the same paths you are GOING to walk down. Vendors should be able to share existing customer names and contact details or arrange for discussions with these customers on request.

Consider each vendor’s recommended platform, architecture, and integration capability. Will the new solution be able to be integrated with the other systems that you have in place today? Can the vendor cost-effectively deploy it at the scale of your IT structure? How do they handle many users, many assets, and many data centers? How do your geographically dispersed data centers affect the DCIM suite’s performance and real-time monitoring capability?

Look for vendors that can provide NEW levels of visibility and analysis, in ways previously not available to you. You are not looking for a prettier way of doing what you can already do, you are looking for new business management insight to allow yourself to make more informed decisions, respond more quickly, etc. Visibility down to the device level is just the start of a solid selection, and what the vendor’s offering does with that level of granular information is where the magic comes from.

Once you have a short list of vendors, you should require out-of-the-box, demonstrable capabilities. The DCIM marketplace is relatively new, and nearly all vendors want to please prospects. When looking for a DCIM vendor, you really want to consider which of these capabilities they can deliver today, and avoid the more theoretical discussion about what they could do given enough time and enough money. Engineering projects create orphaned installations, and can diverge so much from any commercial offering, that customers will be abandoned at the onset, and will not be able to take advantage of the selected vendor’s future releases of their software. As a rule of thumb, if they can’t show you specific DCIM features, they probably don’t have them built yet. Be cautious here as this will determine your long-term costs for DCIM.

Last, ask about pricing models. Be specific. Software vendors are notorious about turning what was presented as a product into a time and materials project. This can be a costly approach. Choose a mature DCIM vendor that details their cost structure, which pieces are off the shelf, and which are custom. They must also articulate ongoing maintenance costs.

33.5.2 Considering the Costs of DCIM

This is one of the most misunderstood topics when discussing DCIM, since the definition of DCIM is so diverse. As mentioned earlier, there are a handful of management software suites that comprise the top-level DCIM functionality. This enables the business management aspects of the deployed solution, and it is the most common interface that users will interact with. All of the rest of the offerings in the DCIM space are actually subcomponents of the total solution. Gartner refers to these vendors as “DCIM-Specialists” or simply “enhancements” to the DCIM solution. These enhancements include hardware and/or software that provides real-time data about power, or environments, allows deeper analytics or customer presentations, or even the ability to discover and identify various assets and their locations.

Today, there is no single pricing scheme for DCIM. The DCIM software management suites are priced in a wide range of schemes based on size or capacity, with perpetual and subscription licenses further complicating the process. Although different pricing schemes exist, for comparison purposes we can use a unit of measure at the “rack” or cabinet.

DCIM enhancement components on the other hand are much more straightforward in pricing. Sensor vendors, for instance, can tell you exactly what a thousand sensors would cost; and if you plan to use four sensors per rack, you can do the math to determine what these 250 racks would cost to outfit with a DCIM-Specialist vendor’s sensors.

So, what does it cost when you are looking to budget DCIM for an upcoming project? As a general rule of thumb, based on the value of dollars in 2014 DCIM Suites and their natural DCIM-Specialists (enhancements) should be budgeted at about US$1000 per rack. This will include core DCIM core functionality, basic integrations with common systems, real-time sensors, installation, and training. (Intelligent rack-based PDUs will add another $1000 per rack, if intelligent power metrics are desired.)

33.5.3 Other DCIM Considerations

DCIM solutions are coming of age and are almost at a level of maturity that large and small organizations alike can begin to take advantage of this new area of data center management. It has never been easier or timelier to create an extended view of the data center infrastructure, by extending the logical views already deployed with physical layer extensions found in modern DCIM offerings.

A few points worth considering as you begin to investigate and then formulate your DCIM plans:

• What is your adoption timeframe, or can you afford to do nothing?
• What are your existing sources of truth and other documentation used in production today?
• Who will be the owner of this project and what resource are committed to it?
• Once DCIM is deployed, where do all the existing IT Support people go?
• Acknowledge your DCIM needs and capabilities will evolve over time.

33.6.1 Consolidation and Rationalization of Vendor Solutions

DCIM has been an emerging technology of increasingly high interest since the mid-2000s. Referring back to Gartner’s “hype cycle,” we saw the peak in vendors of any type of DCIM during the 2010 timeframe where more than 100 vendors self-declared their participation in the DCIM marketplace. Partly because DCIM continues to be poorly defined across the industry, and partly because it is viewed as an emerging greenfield for management vendors with few incumbents, the potential DCIM customer has been bombarded by these vendors all making overlapping and, in many cases, unsupportable claims. There has always been a bit of “yellow journalism” afoot in the DCIM space.

It is clear this cannot continue. The value of DCIM is too great to allow the market to continue to be fragmented and confused. Recently, The 451 Group has attempted to provide some guidance into this confusion and have begun to try and define various aspects of the value possible with DCIM by showing how it relates to Service Management. They have introduced the term Data Center Service Optimization (DCSO) which can be thought of as a benefit-oriented superset of DCIM. From their definitions, DCSO systems are used to plan and optimize datacenter resources and services for availability, agility, and financial, operational, and energy efficiencies. Physical and virtual resources include critical systems, assets, power, compute and IT services, and applications.

Other analysts will surely follow this trend and begin getting much more specific about the business value of DCIM solutions rather than focus primarily on product naming.

33.6.2 Automation and Control

Control is a broad topic that will transform DCIM suites from visibility and analysis solutions into well-orchestrated business and workload management solutions that focus on dynamically adjusting ALL resources that are required to meet computing demand. While there are a few functional DCIM systems that offer a level of hardware control across a chosen physical reference design, the promise for DCIM is that true multivendor automation will occur across the entire hardware and software platforms found in the common data center, which will rely on a highly functional DCIM layer itself. DCIM with control plane capabilities will come of age over the next 10 years. While there are a number of startups today that focus on these automated approaches to dynamic capacity management (cooling and processing), the overall market has not yet embraced rallied behind these concepts as mainstream, and there are several active community discussions which describe commercial data centers which are operational in the year 2020 and beyond as being completely self-healing and dynamic in capacity, where automation will precisely align supply and demand for computing, along with all of the physical resources required to do so.

33.6.3 Asset Location, Physical Discovery

Asset location is one of the Holy Grails of the DCIM market. Today, there is no standardized way to determine where an asset is physically. In the logical world, active devices can quite easily be detected and interrogated to determine where they are on the network, the type of device, and the services which are running. In 2014, the Open Compute Project proposed a new rack platform, the first new rack design in more than 25 years. This platform once again does not include the built-in ability to physically identify down to the physical location slot where devices are populated. While the Open Compute industry is still working on their final designs, only time will tell if/when asset location will become a reality in any form-factor. Until then, innovative startups will continue to look for more ways to retrofit standard racks to enable this capability through the use of wired, wireless, and optical interrogation technologies.

33.6.4 Ecosystems and Integration “Standards,” Linkages to Other Systems

The DCIM marketplace is poised for strong partnerships to form. Potential customers are looking for the DCIM vendor community to consider all of the strategic pieces required to demonstrate core data center management value, and then seek out those portions which they do not make themselves. Prospective customers of DCIM are looking for the “heavy lifting” for integrations to be done by the DCIM vendors involved. It is no longer enough to hide behind standard statements that speak about “protocols” such as “SNMP” or “WebAPI” as their sole approach to integration. Experienced IT professionals understand that general purpose support for standard interfaces is a far cry from systems that can seamlessly work together. These potential adopters of DCIM are looking for strong Ecosystems to form.

Reference

1. [1] Uptime Institute. May 2012. Data center annual survey. Available at http://uptimeinstitute.com/2012-survey-results. Accessed on May 27, 2014.

Further Reading

1. Azevedo D, Belady C, Patterson M, Pouchet J. September 2011. Using CUE™ and WUE™ to improve operations in your DataCenter. The Green Grid.
2. Belady C. October 2007. The green grid data center power efficiency metrics: PUE and DCiE. The Green Grid. WP #06. Available at http://www.thegreengrid.org/Global/Content/white-papers/The-Green-Grid-Data-Center-Power-Efficiency-Metrics-PUE-and-DCiE. Accessed on May 27, 2014.
3. Belady C. 2010. Carbon usage effectiveness (CUE): a green grid data center sustainability metric. The Green Grid. WP #32. Available at http://www.thegreengrid.org/~/media/WhitePapers/CarbonUsageEffectivenessWhitePaper20101202.ashx?lang=en. Accessed on May 27, 2014.
4. Blackburn M. January 2010. THE GREEN GRID data center compute efficiency metric: DCcE. The Green Grid. Available at http://www.thegreengrid.org/~/media/WhitePapers/DCcE_White_Paper_Final.pdf?lang=en. Accessed on May 27, 2014.
5. Cappuccio D. March 2010. DCIM: going beyond IT. Gartner ID: G00174769.Gartner Inc., Stamford, CT.
6. Cappuccio D, Cecci H. June 2012. Cost containment and a data center space efficiency metric. Gartner ID: G00235289. Gartner Inc., Stamford, CT.
7. Clark J. October 2011. The price of data center availability. Data Center J. Available at http://www.datacenterjournal.com/design/the-price-of-data-center-availability/. Accessed on May 27, 2014.
8. Clark J. November 2011. The next data center real estate boom. Data Center J. Available at http://www.datacenterjournal.com/facilities/the-next-data-center-real-estate-boom/. Accessed on May 27, 2014.
9. Cole D. June 2011. Data center energy efficiency—looking beyond PUE. No limits Software. Available at http://www.nolimitssoftware.com/docs/DataCenterEnergyEfficiency_LookingBeyond.pdf. Accessed on May 27, 2014.
10. Cole D. May 2012. Data center knowledge guide to data center infrastructure management. No Limits Software. Available at http://www.datacenterknowledge.com/archives /2012/05 /22/guide-data-center-infrastructure-management-dcim/. Accessed on May 27, 2014.
11. Data Centre Specialist Group. May 2012. Data centre fixed to variable energy ratio metric. Available at http://dcsg.bcs.org/data-centre-fixed-variable-energy-ratio-metric-dc-fver. Accessed on May 27, 2014.
12. EPA. June 2012. Annual energy outlook with projections to 2035. DOE/EIA-0383. Available at http://www.eia.gov/forecasts/aeo/pdf/0383(2012).pdf. Accessed on May 27, 2014.
13. Fichera D, Washburn D, Belanger H. November 2012. Voice of the customer: the good, the bad, and the unwieldy of DCIM deployments Forrester Research, Cambridge , MA.
14. Fry C. January 2012. Green data center: myth vs. reality. WWPI. Available at http://www.wwpi.com/index.php?option=com_content&view=article&id=13817:data-center-infrastructure-management-myth-vs-reality&catid=210:ctr-exclusives&Itemid=2701757 (iTracs). Accessed on May 27, 2014.
15. Harris M. June 2012. DCIM value: two halves make the whole!. Available at http://dcimexpert.com/2012/06/two-halves-make-the-whole/. Accessed on May 27, 2014.
16. Harris M. 2009. Taxonomy of data center instrumentation. Mission Critical Magazine Available at http://www.missioncriticalmagazine.com/ext/resources/MC/Home/Files/PDFs/WP-Taxonomy_of_Data_Center_Instrumentation-Mark_Harris.pdf. Accessed on May 27, 2014.
17. Howard C. February 2012. Hybrid IT: how internal and external cloud services are transforming IT. Gartner ID: G00231796 Gartner Inc., Stamford, CT.
18. IBM Global Technology Services. February 2012. Data center operational efficiency best practices. Ref: RLW03007-USEN-01.IBM Corp., Whiteplains, New York..
19. Intel. February 2012 Moore’s law inspires Intel innovation. Intel Corporation. Available at http://www.intel.com/content/www/us/en/silicon-innovations/moores-law-technology.html. Accessed on May 27, 2014.
20. Kaplan J, Forrest W, Kindler N. July 2008. Revolutionizing data center energy efficiency. Available at http://www.ecobaun.com/images/Revolutionizing_Data_Center_Efficiency.pdf. Accessed on May 27, 2014.
21. Kumar R. July 2008. The six triggers for using data center infrastructure management tools. Gartner ID: G00230904. Gartner Inc., Stamford CT.
22. Mell P, Grance T. September 2011. The NIST definition of cloud computing. NIST. Pub #800-145. Available at http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf. Accessed on May 27, 2014.
23. Neaves R. September 2011. Moving the data center from chaos to control. Nlyte Software. Available at http://www.nlyte.com/german/doc_download/15-moving-the-data-center-from-chaos-to-control-white-paper. Accessed on May 27, 2014.
24. Ortiz Z. February 2012. The green grid monthly members webcast: productivity proxies, economizer survey, and forum 2012. Available at http://www.thegreengrid.org/en/events/Forum2012PreviewWebcast.aspx. Accessed on May 27, 2014.
25. Pultz JE. December 2011. Net IT out: Data Center Infrastructure Management (DCIM): new tools to monitor, manage and control power. Las Vegas Conference. Gartner Inc., Stamford, CT.
26. Pultz JE. February 2012. More than half of data center managers polled will likely be using DCIM tools in 2013. Gartner ID: G00231803. Gartner Inc., Stamford, CT.
27. Ravo K. 2012. A new power measurement standard, UL 2640, to reduce data center costs. Available at http://www.ul.com/global/eng/pages/corporate/aboutul/publications/newsletters/hightech/vol2issue3/4par/. Accessed on May 27, 2014.
28. Schreck G. December 2009. Put DCIM into your automation plans. Forrester Research, Cambridge, MA.