This tip is part one of a series on using data center infrastructure management tools. Read part two: using DCIM
tools to monitor data center power.
As the name implies, data center infrastructure management (DCIM) helps organize a company's infrastructure and facilitate data center management. However, if you ask different people in the data center world what "infrastructure" means, you will get diverse answers. The same holds true for the current crop of DCIM offerings by various vendors.
DCIM tools and their role in data center management
In some cases, DCIM products are based on adaptations of traditional building management systems (BMS) platforms. These tools are geared toward providing information on facilities' energy usage so that the facilities team can better manage the major power and cooling systems.
Some DCIM tools are basic systems that gather the fundamental energy information, or the "total IT power," from the output of the UPS and then compare it to the "total facilities power" to give you a power usage effectiveness (PUE) result. This baseline information should be used for improving the data center site's power and cooling efficiency, as well as optimizing data center management.
Other DCIM platforms are aimed at helping IT administrators monitor the power and environment at the rack level. Some vendors' DCIM tools incorporate a mix of IT asset management capabilities and energy usage monitoring. A number of DCIM vendors also offer "dashboards" that provide visibility into the power consumption of individual racks, individual servers or other IT gear. DCIM packages can even provide a graphical map of each rack and its individual IT equipment, as well as its energy utilization, temperature and humidity. Of course, no DCIM package would be complete without the proverbial PUE metric calculator. Therefore, even the IT-oriented packages need to be able to accept information about a facility's power usage.
At this point in time, very few DCIM platforms are used to directly control IT equipment functions. They are also incapable of automatically scaling and directly controlling a facility's power and other equipment, such as chillers, in response to varying computing loads. Most DCIM tools are designed to give better insight and analysis into the energy utilization of different systems. Of course, the necessary energy-measuring equipment must first be purchased and installed.
Using DCIM tools for energy monitoring
The purchase of energy-measuring equipment is a stumbling block for many projects. The required electrical work for installation of the equipment is an additional expense, and in some cases, systems must be shut down before the electrical work can begin. And of course, in the mission-critical world of the data center, "shutdown" is not in the lexicon.
For DCIM systems to gather and analyze energy usage information, energy-monitoring hardware must be installed at the point of utility handoff, at electrical panels and with facilities-related equipment, such as chillers, CRACs, pumps, etc. Installing at these points is necessary for in-depth analysis of the cooling infrastructure.
On the IT side, DCIM systems might be used just for gathering UPS output data (for a top level view), or for providing more detailed information through rack- or IT device-level monitoring. The most common way to measure IT equipment power is by installing intelligent power distribution units (iPDU). This will be discussed in more detail in part three of this series on data center energy monitoring best practices.
Some vendors offer DCIM software that can accept data from other vendors' existing equipment, such as a UPS, as well as dedicated hardware. Other vendors provide hardware sensors and software. Some vendors offer DCIM functionality as a hosted service -- the sensor information is sent over the Internet to hosted systems and is viewed over a Web interface. DCIM software is also available for installation on a local server.
Some DCIM platforms offer the combined functionality of asset and energy management. Of course, in some cases, the platform is only a reporting tool that has no ability to control any infrastructure or IT systems equipment.
DCIM challenges and limitations
One of the major obstacles that DCIM vendors face is making the DCIM package suitable for facilities teams and IT departments. Satisfying each side's opposite expectations is difficult, and the technical and logistical issues associated with the installation of the energy-measuring equipment must be handles by knowledgeable personnel. Moreover, these data center management teams must be able to overcome the potential stability and security issues of interfacing and integrating any DCIM agent software with the IT computing architecture.
A few software-only vendors have combined forces and partnered up with the manufacturers of energy-monitoring hardware sensors to deliver a complete solution. However, every implementation is different, and cooperative efforts between IT and facilities teams are necessary just for successful installation of energy-monitoring hardware. Interfacing and extracting information from an existing building management system also takes combined efforts.
As previously noted, each group or department has different expectations of what DCIM means and the functions it should perform. Facilities teams expect DCIM to be focused on the power and cooling infrastructure. The IT group expects DCIM to reach the rack- and IT-equipment level. Ideally, DCIM should look at the IT processes and computing workload, the applications and virtualization software and the underlying computing hardware to improve data center management.
Ultimately, the most effective DCIM tool would be truly dynamic, able to receive and understand the computing loads on a real-time basis and ramp up (or ramp down) the power and cooling systems to meet the expected loads. Right now, we are only seeing the beginning of DCIM development and a slow cultural acceptance of any proposed control issues.
Another underlying problem is who has ultimate control and responsibility over system management and operations. The IT group does not want the central management console to resemble a facilities-type infrastructure management package, which may affect the operation and stability of a complex IT systems architecture. And the facilities side does not want the IT group's server management software controlling the chiller plant.
Read part two of this series: Using DCIM tools to monitor data center power
About the author
Julius Neudorfer is the CTO and founding principal of NAAT. Neudorfer has designed and managed communications and data systems projects for both commercial clients and government customers.