It's now or never for old sysadmins to learn new tricks
Watch out for your jobs, says Trevor Pott
Sysadmin blog In most fields of human endeavour the complete invalidation of a person's formal training and skillset generally takes decades, if not generations.
Within IT the tools, applications, operating systems and cloud services learned at the beginning of a bachelor's degree can already be defunct before that degree is completed.
I believe that the traditional roles of IT administrators are evaporating and I want to explore the options remaining to them.
We happy few
The current round of IT job evaporation rests on the rapid uptake of software-defined "X". This ranges from virtualisation to clouds to software-defined storage, networking, even entire data centres.
These tools make it easy for a handful of half-way competent generalists to accomplish what once required teams of specialists.
Terrified of being demonised, vendors of software and services have collectively invested ten of millions of dollars in the past few years to convince us that our jobs are not going to disappear.
Enterprise administrators working for companies that stay as far from the bleeding edge as possible loudly agree with them. These people feel their jobs are secure and that it must be true of everyone else.
The word "bloodbath" has been used more than once
Sysadmins who work for small and medium-sized businesses (SMEs) tell me a different tale. The word "bloodbath" has been used more than once; the past few years have seen our numbers dwindle.
Jobs are winking out like lights at dawn and we nervously chatter on social media about the best path forward. Wage erosion and position evaporation starts with us SME admins, but the decimation will spread upmarket soon enough. What can we do?
The tough get going
The concept of quitting while you're ahead needs a thorough look. For some, moving beyond their area of expertise and experience would be the equivalent of starting from scratch.
Those who chose to specialise in an area of IT that is being rapidly commoditised face massive competition as that niche collapses.
This competition will create a downward wage pressure that will ultimately drop the salary people are willing to pay. That once-cushy job may well drop below the red line where you can't live on that wage any more without some very dramatic changes in housing or other lifestyle elements.
Reskilling is not for everyone. As much as some would have you believe you can simply wander on over to Trainsignal, take a few courses and be off on a new career, it doesn't work like that for everyone.
Many people are too stuck in their old ways, their prejudices and biases too ingrained to be reprogrammed for the new order. For these folks exiting IT altogether may well be the best plan.
The perfect blend
The CIO’s job and the administrator’s are slowly merging. The usefulness of the MBA-trained CIO is at an end. In today's world IT managers – indeed managers in general – can function only if they understand both technology and business. The two are intertwined and inseparable but only a few individuals are capable of bridging the gap.
In truth the role is evolving. The CIO has moved beyond being an individual providing oversight of technological operations within a company.
I call the modern CIO replacements "cloudherders." They find the best mix of in-house and hosted technologies and balance them against business needs.
In a smaller business this may mean managing the odd system directly; in larger companies cloudherders evolve almost exclusively into "vendor relations specialists", rather than coalface admins.
The duty of this New World CIO has nothing to do with infrastructure and everything to do with managing the flow of data from a strategic level.
Plumbing the depths
Doubling down on a specialisation is a route that enterprise administrators recommend. I call it gambling. There is a lot of noise from which to extract signals and make out what is reality and what is wishful thinking.
Without question the über-specialists have the possibility of making far more money at their craft than the rest of us. The upper limit on a Cisco Certified Internetwork Expert (CCIE) is sky-high but the median salaries are still generally below $125K. In no way do you need CCIE-level certs to make $125K in IT. The gap is narrowing on the benefits provided by these certs.
If you have a natural talent for (or a lifetime's experience in) a given specialisation then delving even deeper into it might well be a good call. Similarly, if you are confident of either your connections or your charisma you might make out like a bandit.
Outside of those possibilities, I can't wholeheartedly recommend certifications simply because that is what everyone else is recommending. It takes about $20K to pursue a CCIE-level certification with no guarantee that by the time you are done you will be making much more than you are now.
This is where the gambling side of the equation comes in. Pursuing deep certs without some natural wellspring of charisma to back you up is all about picking the cert path that will both be in demand and not flooded with others seeking the same riches and glory. So $20K to end up little better off than you are now seems pretty steep.
Follow the rocky road
Instead of trying to find some niche in IT that hasn't yet been commoditised and hiding out there until they come for that job too, I have decided to embrace the dark side.
Instead of running one shop I run 20. I use software and technologies that allow me to automate the businesses and charge a small retainer to monitor the systems and be there in case of emergency.
There is a lot of risk in becoming a managed service provider. It means going into business for yourself and that is always a rough road. Yet if you have a desire to do more than hold your financial ground this may be the only realistic path for many admins.
When one SME can no longer be expected to support your salary in a software-defined, cloud as-a-service world, you have to look to multiple companies to fill that function.
If you aren't already doing DevOps, then you are probably dead and you just don't know it yet
The closest thing to a true evolutionary path for sysadmins is the transition to DevOps. If your company qualifies as commercial midmarket and you aren't already doing DevOps, then you are probably dead and you just don't know it yet.
I work with companies with as few as 15 staff that have custom, in-house software. Many others need to beat software-as-a-service (SaaS) solutions into some semblance of utility, often involving building custom parsers (which usually evolve into middleware stacks) that pull information out of one SaaS app and insert it into another.
Even as I see these smaller companies pulling away from managing their own in-house infrastructure, I see their demand increase for software tools that bind various bits of infrastructure together into a cohesive whole.
It is not quite development, it is not quite operations, but it is in demand.
Find the exit
Positioned at the heart of the DevOps movement, Puppet has done a lot of research on the subject. The company has done surveys for years and certainly talked to people from all over the industry.
At first blush I found some results hard to swallow, but it is hard to poke holes in its methodology. After nearly an hour on the phone with their survey team the best I managed was to convince them to add a couple questions for next year.
The future of IT is pretty clear: we won't be administering systems as we do today. Configurations will be dynamic, states imposed based on responses rather than pre-determined conditions.
We are still a long way from that for most places – but the world changes fast. It takes years to reskill an individual and in those years our entire industry will have reinvented itself at least once.
Traditional administration jobs will become fewer and farther between. Picking an exit path is looking more important with each passing month.
Which path will you choose? Why? What options have I missed? The comments section is below; you know what to do. ®