When a new market segment starts to emerge, some analyst company
tends to name it. The data center infrastructure management (DCIM coined by Gartner) segment is now emerging in the data center space.
DCIM solutions collect data from both the IT and the facility parts
of a data center. I am familiar with companies like Sentilla, Modius, OSIsoft, and
SynapSense. Arch Rock was spun off from Cisco and spun back in recently. Power
Assure provides somewhat more sophisticated power management for data centers.
Those DCIM companies collect real-time data from actual operations
and provide varying degrees of functions. Some collect data from both IT and
facility equipment (like servers), aggregate it, and display the result to
provide an overview of a data center’s power usage. Others receive data from
somewhere else and provide more sophisticated analysis.
Romonet was founded by Zahl Limbuwala (CEO) and Liam Newcombe back
in 2006, but they kept it in stealth mode until now. In conjunction with the
recent DatacenterDynamics conference in San Francisco, Romonet came out of stealth
mode and launched in the US. It launched in the UK late last year.
Wanting to coin a new term to accurately describe their segment, they
came up with data center predictive model (DCPM). Rather than collecting real-time
data, they predict data center configuration and architecture.
They showed the differences between DCIM and DCPM in the following
Romonet’s product is called Prognose. Its function is summarized in
the following slide.
The tool can provide "what if” scenarios for many different
elements, such as PUE and power consumption. Two screenshots are shown below.
This display shows how PUE might change with different power loads
This display shows the power usage information of different IT
The rationale for a tool like this is the complex interrelationship
of elements in a data center. Changing one element may have an adverse effect on
other elements. It would be nice if we could tell what the impact of a change might
be before we make it. Prognose can be used for capacity planning. One of the
case studies presented at the launch meeting was from Intel. A representative
from Intel said that this tool could be used for choosing a data center
location on the basis of temperature and humidity conditions in each
geographical area in the world.
The tool is based on modeling algorithms, and its effectiveness depends
solely on how good such modeling is. They surveyed many data centers of various
sizes to fine-tune the model. Because I have not used this tool for a real data
center, I withhold my judgment on it, but a tool like this is pretty handy when
a data center goes through frequent changes, as they typically do.
Another area where I withhold my opinion is their claim of "only one
DCPM in the world.” This is because I found Nlyte Software
http://www.nlyte.com/ at the show the next day. Nlyte also provides predictive
modeling. They also provide management and real-time monitoring of data center assets.
Claiming differentiation by just monitoring, aggregating, and
displaying data from multiple sources at a data center is difficult. The
differentiation is in the analytics and prediction. As Romonet said, the DCIM
segment is crowded, and some consolidation is inevitable. It is not "if” but