Kit Wai Chan - Fotolia
Good DCIM software should be a valuable tool that helps manage operations and alerts to impending problems -- with as little user involvement as possible. It may provide much more, but only if you can handle it.
All data center infrastructure monitoring and management (DCIM) tools have one thing in common: data. But a DCIM system that tells you everything about that data may be sharing more than you need to know.
No one can actually define DCIM or explain precisely what it's supposed to do. Why? Every tool is different, and the data center infrastructures they monitor vary widely. Any information beyond essential alerts is far-ranging, diverse and detailed. It really suits operations with staff support that can make use of the input. Data centers are complex and they all could benefit from a DCIM tool, but only if the right tool is selected.
Before DCIM software and hardware, data center operators were happy just knowing if the temperature was rising, a cooling unit had failed or a UPS had gone into bypass. Now, servers, air conditioners, UPS systems, chiller plants, power strips and all other equipment spit out data on every facet of their inner workings. It's important to better manage our data centers for energy efficiency and reliability, but equipment designers apparently think that, since 256 is a nice binary number, they have to use every available bit for a data point -- whether it's important or not. It would be easy to conclude that 'too much information' aptly describes the world of DCIM.
Or does it? We can easily get too much data for any person to digest, but can we ever really have too much information? There's a big difference between data and information. Data is every transaction, while information is a graphical representation of data points, with some guidance on points of interest such as a red flag.
Your DCIM tool should accomplish the same thing. Packages that capture more than basic data, and include features like IT asset tracking, can certainly be useful, but those should not be the primary sales points or reasons for making a purchase decision. A good DCIM offering should turn the mountains of performance data into real management information that betters these complex and dynamic operations. What really defines good DCIM tools is how that goal is accomplished, and that's what you should look for when evaluating a tool.
Simple DCIM tools keep it secure
Data center staff often ask if DCIM can minimize the need to work with facilities and building management teams. That should not be a goal. In fact, a good DCIM implementation should help bridge any gaps between IT and facilities -- but without inundating the facilities staff with useless data. The same goes for building management. A good DCIM tool feeds selected alarms and alerts to the building management or automation system, which monitors the whole campus, and often connects to the security desk. These people deal with much more than the data center and have minimal training on its use and operations.
The right data center management information
If the most important task of a DCIM tool is not to gather data, but rather to condense it into useful information, what makes the information most useful?
Useful monitoring output varies from one data center to another, and can depend heavily on what needs to be monitored and what data is actually available. Also, consider who uses and maintains the tool to make informed DCIM decisions.
Every DCIM product offers standardized graphical user interfaces (GUIs). The main screens should be logical enough for any user to quickly grasp at least 85% of what's shown without further explanation. If you can't do that, it's not intuitively designed and it's not the GUI you want. In an emergency, nearly anyone should be able to quickly zero in on what's wrong, even if they haven't used the DCIM system in a while or had detailed training on it. A good DCIM tool should also generate easy-to-understand trends based on all that data, and ideally will also provide alerts before critical systems actually fail.
Maintaining the deployment should not be a full-time job. Staffing is tight enough in data centers; if it takes too much time to use and maintain a DCIM, the investment proves to be of minimal value. Maintenance time is partly a function of the DCIM design, but it can also result from overbuying a system.
A robust DCIM package should be modular. Get only what you need and can really use at the outset. Once you're comfortable with the DCIM decision, evaluate the utility, cost and maintenance requirements of other modules to add.
About the author:
Robert McFarlane is a principal in charge of data center design at Shen Milsom and Wilke LLC, with more than 35 years of experience. An expert in data center power and cooling, he helped pioneer building cable design and is a corresponding member of ASHRAE TC9.9. McFarlane also teaches at Marist College's Institute for Data Center Professionals.
Tips for choosing DCIM tools
Top 5 challenges hindering DCIM adoption
The reality about DCIM vendors' promises