This is Part 3 of our five-part series on the countless number of decisions an organization needs to make as it embarks on the DCIM purchase, implementation, and operation journey. The series is produced for the Data Center Knowledge DCIM InfoCenter.
In Part 1 we gave an overview of the promises, the challenges, and the politics of DCIM. Read Part 1 here.
In Part 2 we described the key considerations an organization should keep in mind before starting the process of selecting a DCIM solution. Read Part 2 here.
In the first two parts of this series we examined the vendor promises for DCIM and guidelines for developing an RFI or an RFP. In order to maximize the potential benefits, each organization needs to define its own objectives. Therefore, you should be prepared to do a self-evaluation. How well is your organization currently managing its own data center and IT resources. This will help emphasize which areas require the greatest improvement and then focus on what aspects and features of the myriad of available DCIM tools are most needed to address these issues.
Potential Facility Benefits
Some of the potential facility benefits are direct and more readily quantified. The most well-promoted and well-understood is measurement and improvement of facility energy efficiency, expressed as PUE, or Power Usage Effectiveness. DCIM in and of itself cannot fix a major fundamental problem. However, it can provide a deeper view into the individual elements that collectively consume energy (electrical and cooling systems), but in many older facilities they are not monitored separately. For example, in a large older facility with relatively poor energy efficiency (i.e. a PUE of 2.5), the largest energy waste will typically be the cooling system. There are some basic practices that can help improve cooling efficiency (such as just raising the temperature a few degrees and installing blanking panels in the racks) without investing in a DCIM system. In many cases some simple changes may reduce the cooling system energy usage (and cost) by 5-10 percent.
While simple changes are cost-effective, further improvements can be more elusive. One of the direct benefits of a DCIM system is that it can be used to measure and track all cooling system components (CRAC, CRAH, chillers, pumps, etc.) individually while correlating that data to the IT heat load. This allows it to display how much a change in one sub-system can impact the operation and performance of individual components and overall energy efficiency. This can result in a more substantial reduction in cooling energy, which can save recurring energy costs, or one element of overall ROI justification for DCIM – a topic we’ll address in detail in Part 5 of the series. However, if the site already is relatively efficient (i.e. a PUE of 1.5 or less), it is less likely that DCIM will be able to provide as much energy savings, which will reduce the value of that aspect of a DCIM system.
Regardless of how much energy is saved, by implementing row- or rack-level environmental sensors within the whitespace, DCIM can help ensure that when temperatures are raised, there are no racks that are running too hot (i.e. outside of ASHRAE envelopes). It can also display any “hot spots” (typically due to improper airflow issues). It can then measure the potential benefit of any remediation efforts which can help optimize them by monitoring the results as they are deployed. This rack- or row-level monitoring has an indirect but very significant benefit of potentially improving IT equipment reliability by helping maintain ASHRAE-specified environmental conditions, especially in higher-density data centers.
Depending on the capabilities of the DCIM vendor’s offerings, Computational Fluid Dynamics-type modeling (or actual CFD capabilities) can provide graphic thermal mapping functions. This feature can also calculate and display “what-if” airflow scenarios for new equipment placement, which is also an element of overall capacity management. This also allows for cooling system impact zone failure analysis. For example, if a CRAC unit fails or when it needs to be shut down for maintenance. This helps improve overall availability, which has indirect, but very valuable benefits. CFD functionality, however, can be a relatively expensive option. You may want to consider this as an add-on feature, if it can be purchased later on, once the basic system has been successfully installed.
In many cases the Building Management System has some existing sensors connected to it, which can provide information to the DCIM system. This requires a hardware gateway to interface with the existing BMS. The implementation usually involves system integration services, which typically requires cooperation of the BMS vendor and costs to that vendor. From a practical perspective, if the existing BMS vendor is proposing their own DCIM package, they should be in the position to offer a low-cost (or no-cost) BMS integration as part of the enticement to purchase their DCIM system. This also presumably would avoid multi-vendor finger pointing. However, if they are not the selected DCIM vendor, you should anticipate system integration costs to become significant, since the BMS vendor will have no economic incentive to unlock their proprietary system.
There are obviously several major cost areas associated with the above simplified facility example, starting with the basic cost of the core DCIM software and any optional feature modules. Furthermore, many vendors have additional license costs, which are typically based on the number of monitored devices (per electrical circuit, for example). In addition, there is the cost for a substantial number of environmental sensors for row or rack monitoring of the whitespace, as well as their installation. Finally, there is the cost of the dedicated DCIM server (or servers) that the software is installed on.
Potential IT Systems Benefits
The facility aspect is the most mature and developed area of DCIM. While IT capabilities, features, and benefits have been available for a while, they have been lagging for several reasons. First and foremost, there is still a major cultural divide between facilities and IT. Furthermore, IT systems have many more functional domains, system administrators, and in many cases diverse and proprietary management consoles that are closely tied to each system that may or may not be interoperable with a centralized IT system management console. Moreover, IT software is in a constant state of change, while hardware is also evolving, albeit at a slower pace. Even well before DCIM, many major (and minor) software vendors were trying to be the “uber” centralized IT management console.
On the practical level, most DCIM implementations try to provide a granular look at the energy used by IT hardware. In fact, virtually all modern IT equipment has onboard management systems that can be queried via SNMP to allow DCIM packages to provide real-time power use and environmental conditions. However, despite the fact that this is technically possible, in most cases power monitoring at the rack level is gathered by branch circuit monitoring at floor-level PDUs, or by metered PDUs within the rack (intelligent power strips, some of which can meter per outlet to track energy used by the IT device). One of the main benefits of this is better rack-level capacity management and planning. It also significantly improves availability by helping to avoid circuit overloads and tripped circuit breakers if additional servers are blindly added to a cabinet. This is an indirect, yet very tangible and important benefit.
Nonetheless, there is movement and progress to try to bring the holistic promise of DCIM to fruition, despite, and perhaps because of the shift to the Software Defined Data Center, Software Defined Everything, and of course cloud-based resources. This will be driven by vendors and end-users, as well as service providers. In fact, late last year The Green Grid formed a DCIM work group to help examine the elements and various aspects and benefits of a unified DCIM system. However, IT security concerns represent a substantial impediment to widespread IT adoption of DCIM.
While vendors provide costs in their proposals, there are many indirect costs attributable to your own organization’s internal resources, such as overall project management during the installation process, system administration, and training. These costs should not be underestimated.
One of the promised benefits of DCIM is to improve productivity, presumably by reducing the manual-process workload on staff by automatic polling of data from a multitude of devices. While this is a core function of every DCIM platform, it still needs a system administrator to handle tasks such as user access rights (permission levels, user screen views, etc.), report generation, adding new monitoring devices and system database back-up, just to name a few items. Like any other critical system, other technical staff should also be trained to provide extended coverage hours and back-up for the primary administrator. While overall, if properly implemented, a DCIM system can substantially reduce the workload of manual tasks related to taking and logging measurements, especially in larger facilities (and multiple sites), the added administrative burden should be taken into account.
While the DCIM dashboard should provide a unified view and simplify monitoring and management of all the elements of the data center, like any other system, user training is required to make the most effective use of the system. Some DCIM systems have more intuitive interfaces than others, while some of this will become apparent during a product demo, which will impact how much or how little user (and administrator) training will be required. This may be an added cost from the vendor (the first basic session may be included), however, additional sessions for users, as well as higher-level technical administrator training will be required, perhaps with additional costs. Moreover, the training time allocated by the end-user staff is also an indirect cost which should be considered as well.
The Bottom Line
The above examples only represent a fraction of potential benefits of DCIM. However, the size of your data center(s) and IT systems will also obviously impact the relative value of potential benefits, as well as the overall costs. When crafting your basic requirements or creating your ultimate DCIM “wish list,” plan for the end goal, a comprehensive DCIM system that improves operational and energy efficiency, as well as overall data center availability. But be realistic in your projected timelines and expectations, and in order to be successful, expect to execute in phases. Consider doing a proof-of-concept baseline project first to gain experience and then use that for more accurate budgeting and selection of the most practical functions for a full-scale deployment.
What are the challenges in DCIM implementation? How do you justify the price tag of DCIM? Come back to read parts four and five of this series in the Data Center Knowledge DCIM InfoCenter.