Do Not Tighten Bolts

nuts-and-bolts.jpgI have a pretty incredible job.   My current role has me leading Microsoft’s efforts at designing, constructing, and operating its world-wide Data Center infrastructure in support of our cloud services initiatives, or more correctly “Software + Services”.  

There aren’t many people around tasked with this kind of challenge, in fact the number of companies attempting this challenge can be counted on one or two hands.   One routine question I get asked is ‘what methodology or approach I use to deliver against this challenge?’.  The question of course assumes there is an answer.  As if some there is some book one can run out and purchase to figure it out.   No such book exists. 

The real answer involves the hard work and dedication of an incredibly talented team focused on a single mission.   That is exactly what we have on the Data Center challenge at Microsoft.  I have the world’s most talented team and I am not in the least shy about my confidence in them.   However, even with the raw materials produced by this incredible team,  the technology breakthroughs in Data Center design, the incredible process and automation improvements, the focused reduction of energy consumption and drive for greater energy-use efficiency is not 100% of the formula.

There is one missing element.  Clues to this one missing element can be found in products for sale at your local Target or Walmart store.   Many people may be surprised by this, but I suspect most wont. 

I recently spent a good part of a weekend putting together deck furniture for my home.   It was good quality stuff, it had the required parts and hardware and not unlike other do-it-yourself furniture it had directions that left a lot to be desired. In many ways its like IT Infrastructure or running any IT shop.   You have all the tools, you have all the raw components, but how you put it all together is where the real magic happens, and the directions are usually just as vague on how to do it.

One of the common themes across all steps of the deck furniture pieces was a common refrain, ‘Do Not Tighten Bolts”.   The purpose was to get all of the components together, even if a bit loose, to ensure you had the right shape, all components were in the right place, and then and only then do you tighten the bolts.

If you really want to know the secret to putting together solutions at scale, remember the “Do Not Tighten Bolts” methodology.   Assemble to raw components, ensure you have the right shape and that all components are in the right place, and then “Tighten it down.”   This can be and is an iterative process.   Keep working to get that right shape.  Keep working to find the right component configuration.  Tighten bolts.    As I built my first deck chair, there was significant amounts of trial and error.  The second deck chair however was seamless, even with the same cruddy directions.   Once you learn to ‘Not Tighten’ technique the assembly process is quick and provides you with great learnings.   

Some may feel this approach too simplistic, or lacks the refinement of a polished methodology.   The fact of the matter is that Cloud Services infrastructure takes lots of hard work, great natural resources, and above all flexibility without adherence to dogmatic approach.

That is why I have named this blog “Loose Bolts”.  I will be moving my formal blog activity to this forum and hopefully post interesting topics from time to time.

Thanks so much for reading,

Mike Manos

Author: mmanos

Infrastructure at Scale Technologist and Cloud Aficionado.

3 thoughts on “Do Not Tighten Bolts”

  1. Mike –

    The ‘bolts’ analogy is interesting, and potentially useful.

    Connecting the disparate disciplines of desk chair assembly and IT infrastructure is more interesting.

    I’ve been convinced for quite some time that to make a large leap forward in a technical area, one must look around for similar problems in other non-related fields and see how the problems were solved in that area. For example, what do transportation engineers or manufacturing engineers know that can help IT? Queuing theory? Just-in-time assembly?


  2. Hi Mike,

    Interesting thoughts on the blog. I think leaving the bolts loose must be differentiated from tightening the bolts and throwing out the end product when it’s time to move on. Everything has a useful life, but we do not always know how long the useful life will be. Independent of this knowledge, at some point the cord must be cut, the bolts tightened, and you have to go forward.

    The key about bolt tightening, at least to me, is more of a timing discussion. Not tightening a bolt may lead to more rework than leaving it loose. Unless one possesses tremendous foundational knowledge, making informed judgments on such things is tough.

    Innovation, for example, almost always comes slowly at first, then a torrent of progress follows. If you are caught in the midst of a major change and keep investing the yesterday’s approaches, things will be difficult. Unless of course, today’s approach breaks physical laws or strong social/societal ethos or mores, in which case you are better off investing in yesterday’s approach.


  3. Pingback: Loosely Connected

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: