Where do metrics end and intelligent humans begin?

This audio was created using Microsoft Azure Speech Services

I’m all for metrics, but at some point the need for an actual breathing human to interpret the metric is as important as the metric itself.  PUE, WUE, ERE, DCcE [1] – I’m mentally exhausted –   I don’t think it’s possible to have more letter ‘E’s in a list of metrics.  While PUE certainly has been instrumental in getting the focus on physical infrastructure efficiency effectiveness, I am beginning to wonder if we are heading toward metric overkill.

Most of us know there are limits to PUE if not taken in context.  For instance, a data center consolidation project may increase PUE but can reduce energy consumption.  An intelligent human would say this is a good thing, a metric zombie would say that it’s bad.  So how many metrics do we need and do we run the risk of creating a plethora of metric zombies?  (There seems to be a joke about a CEO in here somewhere but I will refrain in the interest of job security).

WUE – we seem to be moving pretty quickly to reducing the use of water in the data center anyway.  Is this one necessary?  I understand there is this new breed of senior executives called “Sustainability Officers” and the data center manager may need a metric to feed these potential metric zombies.  That being said,  if water is plentiful in the geographic area, does anyone really think a Data Center manager will have this metric high on his priority list?  If the data center is in the middle of the desert, aren’t they already doing what they can to minimize water usage?

DCcE – this one seems interesting but as I thought about it more I’m not so sure.  At the end of the day, most people I talk to would be more interested in knowing whether the cost of running the data center including capital and operating expenses is on par or better than their competitors.  Doesn’t the ‘opex/capex’  metric capture this issue?  If I can write an application that’s more efficient than my competitor which requires less computing power and hence smaller data centers, I will be more competitive since my data center will cost less to capitalize and operate – correct?  Won’t the natural competitive forces make this happen anyway?

And what’s the point of this metric in the context of different businesses – for instance, I’m sure you could compare McDonald’s infrastructure against BankofAmerica but with all due respect – WHO CARES?  Last time I checked they are in completely different businesses so the cost to compute for BankofAmerica may be quite a bit higher than McDonald’s – and probably should be.

Most data center managers I meet are under more fundamental metric –  how green are they.  And I mean green in the context of “Fast Eddie” Felson  (The Color of Money) not Greenpeace.  They really are under pressure to answer one question:  “How much does it cost, and how can I do it cheaper while maintaining my service level.”

I do think metrics can help data center managers achieve this goal, so at the end of the day some will be necessary to help identify opportunities for improvement.  However, let’s not pretend any metric (or collection of metrics with lots of ‘E’s in them) will be perfect.  I think I may start making some t-shirts for data center managers that say:

“I am an intelligent human”

[1] PUE (Power Usage Effectiveness), WUE (Water Usage Effectiveness), ERE (Energy Reuse Effectiveness), DCcE (Data Center compute Efficiency)

[in the interest of full disclosure, my editors’ made me put this footnote in – I figured most of you were would know these anyway so I apologize if I insult your intelligence by footnoting…..]


Please follow me on Twitter @KevinBrown77


About Kevin Brown:

Kevin Brown is Senior Vice President of Innovation and Chief Technology Officer for the €3.7 billion IT Division at Schneider Electric. In this role, he is responsible for driving innovation and managing the R&D portfolio for the IT Division as well as driving the overall Schneider Electric portfolio strategy for the Data Center market. Prior to this position Kevin served as Vice President, Data Center Global Strategy and Technology. Kevin has also held numerous senior management roles in product development, engineering, and software development in the power electronics and HVAC industries. He holds a Bachelor’s of Science in Mechanical Engineering from Cornell University.



Tags: , ,