New research has come out from the Intel DCM's State of the Data Center commissioned report conducted with 200 data center managers across the US and UK. The report revealed some interesting data points. One of the most
interesting findings was that 40% of international data
centers are still relying on manual processes - spreadsheets/physically
walking the data center floor with a tape measure - in an attempt to
"accurately" accomplish capacity planning and forecasting. Perhaps even more
shocking was that data center managers were found to be wasting 40-60% of their time performing manual processes versus leveraging an
automated DCM solution.
To find out more about this report and those findings, I spoke with Jeff Klaus, GM of Data Center Solutions at Intel DCM.
VMblog: What surprised you most
about the fact that 40% of data center managers still rely on manual
processes?
Jeff Klaus: We believed there were still manual
activities happening in the DC, but 40% was higher than anticipated. This
reinforced our mission to continue educating operators about the capabilities
within the hardware they've already purchased! Intel Data Center Manager can
extract data from instrumentation in rack servers and blades for the past two
chipset generations, we need to do a better job educating our customers.
VMblog: If manual processes waste
40-60% of a data center manager's time, how would that percentage shift if
they were to employ a DCIM solution?
Klaus: We have several DCIM use cases, which provide
a range of savings. Some attribute $50-70k savings over a year for a 3MW, 300
rack DC, other use cases save more. It is hard to give a specific number, given
the wide differences in scale and processes that companies maintain. However,
with any given tool there is a good chance that the IT manager can automate and
monitor all servers in real time, gain time savings replacing manual inventory
with automatic data collection, eliminating data collection by ‘walking about'
and the consolidation of underutilized servers. With this in mind, a
15-20% operations cost improvement would seem quite reasonable.
VMblog: And what does this imply about
DCIM solutions/implementations?
Klaus: It seems that there is an ongoing need to
increase awareness with end users of solutions, so that end user know that they
can really solve issues. It could also imply that products need
improvements in ease of use and/or to be more integrated to provide the vaunted
‘single pane of glass' often requested.
VMblog: What's the biggest
challenge faced by data center managers/operators today?
Klaus: There are multiple pressing issues for
IT. There is so much new information, capabilities and promises coming at
them. The industry challenge is to simplify the communication based on some
extreme complexity and let the operators judge the best fit for their
environment. Much of IT's time seems to be in facing the imminent problems of "here
and now". The time to plan for operational efficiency may be a luxury
that is not often available. However, if some time is taken to look at
the available tools that may save themselves plenty of time and money over the
long term.
VMblog: What should data center
managers and operators be thinking about when reviewing/allocating their
2016 budget as it relates to DCIM solutions?
Klaus: There are many solutions on the market and
there are many types of capabilities and usages that are
supported. One way to do this, is for an IT manager to evaluate based on
what they want to achieve in terms of key performance indicators KPI, e.g.,
increase total performance by 20%, but maintain the same operational budget or
reduce overall operational expense by 20%. They should trial
solutions and evaluate relative to these goals to assess the value of using
them.
VMblog: Based on the report, what do
you predict to happen over the coming years? (e.g., manual processes to
stay the same, more widespread adoption, the market will mature as more
DCMs become educated on the benefits of DCIM, etc.)
Klaus: The level of instrumentation in IT devices
will continue to improve and broaden. IoT has come to the data center and it
will continue to accelerate. DC Operators need to get ahead of this infusion of
data and select the best strategy to harness this information and analyze it
for their environment. This implies that higher adoption of tools and processes
that more efficient is inevitable over time. While a few may hold onto
their spreadsheets as long as possible, I would expect the vast majority to
find higher order efficiency becomes an executive management requirement and
the use of DCIM becomes pervasive.
VMblog: Any other surprises from the
report?
Klaus: There is an overall progression that IT
typically goes through as they move from ad hoc processes to higher levels and
scales of efficiency. It has been described before including the The Green
Grid's Data Center Maturity Model, as an example. This occurs over time with
the increase in the expertise at the customer and with the broader availability
of tools to manage the environment. If anything, it is a surprise to see
that many are still using quite basic methods to control their environment,
when there are multiple excellent tools available for use.
##
Once again, a special thank you to Jeff Klaus, GM of Data Center Solutions at Intel DCM for taking time out to speak with VMblog. Looking for more information? I invite you to check out the report's findings for yourself, here.