1. Business-wise, Future-drivenTM
Data Centers
The Data Center Chaos Remediation Tip Sheet
There are six main causes of chaos and instability in the data center, and the good news is that they can be remediated to
improve data center efficiency and reduce downtime. This tip sheet will help you understand the problems, the solutions
and how you can make the changes that will have real benefit.
Understanding the Cause Helps Drive the Cure
The key reason for problems and chaotic infrastructure in the data center is too many years of unplanned, short-term
focused change to the data center. As individual problems are solved without regard to a holistic plan, decisions and
actions are made that drive chaos.
What specifically does chaos look like and what causes it?
1. Inconsistent cabling with no color standards and random routing – this makes it very difficult to diagnose
problems and may eventually lead to downtime when the wrong cable is unplugged.
2. Lack of a documented change management process – while change management is common practice in IT, its
scope rarely includes changes to the physical infrastructure where the impact of downtime can be more widespread
than, say, one application.
3. Unnecessary and overbuilt UPSs – In the fear that some equipment may be unprotected, UPSs may be added
without a clear plan and consistent hardware coverage.
4. Uncalculated reactions to hot spots – when hot spots are found, some operations managers react by reducing
the temperature set points on cooling units, which often creates more problems than it fixes. Turning set points down
spontaneously pulls moisture out of the air, reduces the cooling unit’s capacity, and increases energy consumption to
re-humidify the air.
5. Air flow issues and inconsistent use of blanking panels – airflow problems compound as the data center grows,
leading to a higher power bill and higher risk of downtime.
6. Inconsistent or missing documentation for data center infrastructure – when technicians perform maintenance,
inadequate documentation can lead to mistakes, downtime and other costs.
There are, of course, a number of other attributes that can describe some level of chaos. In fact there have been contests
that showcase some of the more extreme examples of this phenomenon. However, the focus isn’t finding ways to describe
the problem; it’s how to create more order!
The Brands You Trust.
The Brands You Trust.
2. Business-wise, Future-drivenTM
Data Centers
Taking Control – Tips to Resolve Chaos
Tip #1 Organize the Wiring Plant
When the wiring in your server room or data center looks like a modern hair style gone bad, you not only have chaos, but
you also likely have imminent downtime. Start with going to a documented and consistent color scheme where there are
no exceptions and adherence is mandatory. Next, develop a standard wiring routing approach where all cabling is put on
one side of a rack, and ties are used to keep cable bundles neat and out of the way of airflow. Many data center managers
are moving to overhead cabling and power distribution for improved cable management and to maximize air flow under the
raised floor.
Tip #2 Be Serious About Change Management
There is no quicker way to chaos than making undocumented and untracked changes to data center or server room
infrastructure. It’s important to do this within teams, but don’t forget to get out of department “silos” and coordinate between
IT and facilities (and third party vendors!). Use automation tools and document management systems to enforce change
control and ensure all documentation is accounted for and is easily retrievable.
Tip #3 Create a Holistic View of the Data Center or Server Room
Fight chaos by moving from a view of only a single rack, server or device to a holistic view of how the overall datacenter is
impacted by changes. This includes air flow, UPS usage, power distribution, network cabling, and rack/area density. Another
important aspect of this comprehensive view is to ensure that you take into account servers, storage, and networking as a
combined entity, not a piecemeal approach. DCIM tools can help you plan, operate and analyze your data center holistically.
Tip #4 Use Standardized Infrastructure
There is a cogent argument that the single largest cause of chaos in a data center or server room is the use of unique,
customized, and varied physical infrastructure. As different styles of racks have inconsistent cable routing, power locations
and capabilities, IT has to build what amounts to a number of “mini” data centers that all behave differently. It’s easy to
see how this adds complexity to IT operations, but is also causes other problems, like “knowledge walk,” when highly
knowledgeable staff leave the company or change roles. Using a standardized infrastructure provides consistency that not
only reduces the time needed to deal with repairs or issues, but also reduces chaos substantially.
Tip #5 Clean Sheet Contrast
One of the most useful analytic tools is the contrasting of a “Clean Sheet of Paper” of what your data center or server
room might look like, and what it is today. Admittedly it’s an unfair comparison, but the idea is not to look at “what might
have been;” it’s to use the contrast to identify the areas that are most likely creating chaos or other problems. This
approach gives a very different and useful perspective that is hard to have when you’re focused on individual, tactical
problems.
SUMMARY
Chaos in the data center is not just something that makes your day to day life more frustrating, it will cause downtime and
lead to far higher operating costs. While building a new data center is one route to eliminate chaos, it’s not a fiscal reality in
many cases. You can, however, use these five tips to start reducing chaos and improve your operating reality today. Steady
improvement will yield compelling results.
The Brands You Trust.
The Brands You Trust.