Automation is About Enabling People, Not Replacing Them

4 minute read
Luke Kanies avatar

Information Management, Automation is About Enabling People, Not Replacing Them
I spend a lot of time talking with people about IT automation and frequently get the response, “Great, so we can fire people!” Even people who work in IT operations who stand to benefit most from automation draw a straight line between automating and cutting staff -- and they respond with fear.

A Simple Choice

In reality these anticipated firings almost never take place.When someone brings up the idea of reducing IT staff -- especially someone who actually pays for investments in technology, such as a technology executive -- I present a simple choice:Would you rather maintain the same service quality and reduce your cost or would you rather increase your service quality at the same cost?

Not surprisingly, pretty much everyone says their major goal is increasing service quality, not lowering cost -- that is, reducing headcount.It’s just that too many times they clearly see a lever in their financial model that controls cost, but they don’t see the “make things run better and faster” knob in their data center.

Further, too many organizations aren't getting anywhere near the service quality they need from their organization.They see high error rates, slow response times, glacial deployment speeds and almost no information about what’s actually happening and why.

For us, automation is about fixing those service quality issues. Those issues exist not because there aren’t enough IT staff; they exist because IT staff are too often being tasked with things people aren’t good at, which means they don’t have enough time to spend on the things people excel at.

Repetitive Tasks versus Decision Making

If you look at a typical change across 100 hosts, if a human connects to each machine in turn and makes the change, he or she is highly likely to make a mistake. A human is just as likely to learn something in the middle and change the work halfway through.

This is because humans are no good at performing repetitive tasks without making mistakes, or learning, or both.Computers are fantastic at repetition, and I think we can all agree they’re equally good at not learning anything.  Computers generally do not excel at knowing what changes should be made, or when to make them.

Humans are fantastic at pattern recognition and fast decision-making.When you put a ton of data in front of a computer, you have to teach it all kinds of stuff about how to understand the data, and then build heuristics and complicated algorithms to get it to make a half-decent decision. Even then, you usually want a human supervising.Put that same data in front of a human and that person will start recognizing anomalies and impending events almost immediately, and then start making recommendations.

Learning Opportunities

Sysadmins aren’t generally lever pullers; they’re not sitting around performing menial repetitive tasks.Some are though, and these people's jobs actually are in danger, because the value they add can be trivially automated -- and someday it will be.

Eliminating Lever Pulling

The real problem is that sysadmins have to split their time between things they’re good at and trivial lever pulling. It’s a problem that can be solved when you consider that humans and computers have complementary strengths and weaknesses. We want to automate all the lever pulling, all that trivial menial work, so people can focus on the things they’re great at.

This is why we recommend that initial investments in automation should not be large-scale strategic projects. Instead, they should be focused on eliminating the menial, repetitive tasks and mistake-prone processes that require exact performance every time. Automating these areas away gives your people more time and less stress, so they can now build even better automation for more strategic projects.

Sysadmins can start by automating something really simple, such as the motd file, or automating security settings such as the subdoers file; ssh root access; root user password; users and groups; or IP tables and firewall rules. The motivation here is to learn about the automation tool by doing something that’s easy, and realize the immediate benefit of automating away a trivial, repetitive task. As sysadmins get used to their automation tools, they can ramp up and automate more complex repetitive tasks that also take up too much of their time.

So, when you think about investing in automation, make sure you’re taking the strengths and weaknesses of your team into account. Do your best to free people from typing the same command 100 times, and you’ll be enabling them to engage in pattern recognition, fast decision-making and quick responses -- and to add far more value than a computer ever could.

Title image courtesy of karen roach (Shutterstock)

About the author

Luke Kanies

Luke Kanies founded Puppet and Puppet Labs in 2005 out of fear and desperation, with the goal of producing better operations tools and changing how we manage systems. His work with Puppet has been an important part of DevOps and delivering on the promise of cloud computing.