“You cannot manage what you cannot monitor”. This is a common message that has been used in our industry for years and is recently becoming more prevalent in the data center management world.

One of the foundational elements of DCIM (data center infrastructure management) is the ability to unify IT and facilities information using real time data resulting actionable information around subjects like capacity planning, availability, efficiency, asset management, equipment life cycles, redundancy, power, space, and cooling. Industry analysts estimate the current DCIM market at just under $1 billion with as many as 70 vendors selling variations of DCIM software, each with their share of associated pros. and cons. Given the associated complexities, ability to scale, acquisition, implementation, and possible recurring licensing costs, finding the correct solution is a crucial and often daunting task.

You need a Rosetta Stone

At a high level, there are two ways to build a DCIM solution. The first is to build a new system from the ground up by selecting a solution from one of the 70+ DCIM providers. The second is accomplished by merging existing disparate, and sometimes stand-alone systems from multiple manufacturers into a single pane of glass.

What both of these methods have in common is the requirement to integrate existing infrastructure (electrical, mechanical, racks, compute, storage, etc) and make sure you can monitor in real-time. In order to integrate, the DCIM system needs to be a Rosetta Stone of sorts. The system will need to span older and newer devices and protocols, sometimes aggregate multiple protocols and send the information to front-end system.

Most devices are capable of communication, however there may not be a common type of communication between them. Prior to any type of DCIM software selection and deployment, information needs to be gathered on the quantity of devices, their ability to communicate, type of communication, and requirements of ancillary communication devices such as communication cards, protocol converters, communications medium (cabling), and/or dry contact relay replicators.

Some equipment is able to communicate with the addition of a communications card. The type of card and protocols will need to be discovered, procured, and deployed. Often the language and the medium between the devices are uncommon. In other situations, the only form of communication is a “dry contact” (voltage on/off).  These forms of communications have to be aggregated and converted to a common protocol form.

Ancillary sensors may be required to augment existing equipment depending on the type and resolution of information desired. It is imperative to gain an understanding of the quantity and placement of the sensors as they pertain to the desired information collection. Most often the sensor consists of:

• Current transducers to read power consumption throughout the power distribution
• Temperature/humidity sensors on server racks, air handlers, and cooling capacity devices
• Radio frequency identification tags on most or all assets
• Leak detection around susceptible areas.

Assessing your infrastructure

A successful DCIM system is a solution that provides you with truly actionable information which omits disparate information and nuisance alarms. With so many points of information coming in, how do you know which to pay attention to? What should the alerting thresholds be? Is a single point of information good enough or should it be a variable to be used larger formula?

Experience is everything when it comes to ensuring a successful DCIM deployment and thus the importance of a DCIM Readiness Assessment becomes crucial. Whether its conducted with internal resources or through a trusted advisor or partner, the Readiness Assessment is a valuable exercise and will yield important results that enable you to develop a solid project plan in advance of the actual DCIM software deployment.

Best practice steps that a DCIM readiness assessment should include: 

1. Devise a budget for the project.
2. Involve all stake holders and build a consensus on key elements (needs, wants, and nice-to-haves)
3. Build a list of specific key points of information, indicators, and trends. Arrange them in order of priority
4. List monitoring and alerting formulas and the associated sensing points
5. Survey the target equipment, its ability to communicate, the protocol, and medium of communication
6. Perform a gap analysis of required information not provided by existing infrastructure and review ancillary sensors to fill the gaps
7. Narrow down the DCIM solution with the best requirements fit.
8. Review the list of protocols, medium, and ancillary sensors
9. Review the server requirements of the DCIM software
10. Draw a comprehensive schematic of the DCIM network.

Upon completion of the Readiness Assessment you will have identified the key “holes” in your infrastructure with regard to communications needs and other requirements for your DCIM software and you can develop the necessary project plans to enable your infrastructure to communicate with the font-end DCIM software you are selecting.  The time investment in working through these initial steps will pay off in the long run and will go a long way to ensuring a smooth DCIM implementation.

Robert Huttemann is senior vice president and Anthony Gennello is service engineer at DCiM Solutions