IT Ops Career 2015: Where to Point Your Eyes

In my new role at Pluralsight, I’ve been spending a ton of time looking at IT Operations, broadly, as an industry. Part of my job is to figure out where the industry is going, and help get our authors lined up to create effective training in those areas. But I can also share those areas with you… so that you know where you should be looking over the course of 2015.

Now, you need to have the right attitude about this list: it’s stuff to be looking at. These will be a big deal in the future; some near, some far, but none “right now.” So I’m not saying you have to run out and become an expert. I’m saying these are areas that will be in your future.

“But I’m reading this list, and my employer will never do this stuff.” Well, then you should get your resume up to date. Because this list is basically a complete list of the major skills sets that will comprise IT Operations in ten years. Meaning, it’ll be this stuff… and nothing else. So if you’re not going to start learning this on the job in the next few years, you’d better hope your current job will last forever, because you’ll basically be getting outdated and unhireable. Again – over the next decade, not instantly.

And by the way, I’ve decided to consistently refer to the field as “IT Operations” (as opposed to, say, “IT Software Development”) to indicate the operational side of business information technology. This has no connection to consumer spaces. Those of us in the industry should honestly re-adopt “SysOp” as a job title. I miss it.

Here goes.

Big Data. No, really, read this.

IT Operational Analytics (ITOA) is an emerging thing, but it’s a real thing. Imagine the biggest log scraper-and-consolidator you can imagine, sucking historical data off every system in your environment. It then correlates all that information, and provides predictive analysis. “When incoming Web traffic does this, you can expect these servers to get hit like this,” all throughout your infrastructure.

The immediate application of this – and it’s currently a real thing, just early stages – is far superior trend analysis than you’ve ever had, with better ability to correlate trends to business events (like the Spring Clearance Sale that the company just announced, or the article that was published in the Journal yesterday).

The long-term application is to identify triggers (this happened) and instruct your flexible IT infrastructure to respond (migrate these loads to those hosts and prepare to batten down the hatches). The environment responds intelligently based on past patterns, and it does so based on past facts rather than someone’s current intuition.

Don’t just gloss over “big data” every time you read it. Yes, it’s a stupidly overused word. But for ITOA in particular, is has real applicability to your life.


I’ve spent the last what, 8 years preaching automation? I’m done. Automation is over.


PowerShell, for example, was a great automation tool. It’s now moving beyond that into the next logical stage, which is true configuration management. Not what System Center calls configuration management, either. And, regardless of what brand name you like, configuration management is the next big phase for IT. Whether you use Ansible, DSC, Chef, Puppet, Salt, or Pure Black Magic, being able to state what you want your environment look like, and having your environment “make it happen,” is where we’ll all be in a decade.

Yes, the current state of the art for configuration management is still a little raw, and still requires a bit of hackery and fudging. That’s changing mega-rapidly – like, monthly. Remember, a decade ago, Windows XP was new. A decade from now, you won’t recognize modern datacenters. This all plays into how IOTA can help, too: if a monitoring and analysis tool determines that the environment needs to change (based on a trigger), it simply modifies the “desired state” of the environment. The environment then responds accordingly. The monitoring solution doesn’t make changes; it simply orders them, and then environment knows how to “change itself.”

This is massively different from what’s been done in IT before. Ever. You know how everything in IT is basically old stuff come around for another try? Thin computing, thick computing. Distributed, centralized. We go through trends. Desired state configuration (lowercase letters, not necessarily the specific Microsoft approach to it) is new. Our IT governance frameworks – ISO 20000, ITIL, COBIT – don’t even imagine something like this, it’s so startlingly new. And yet it’s very much a thing, right now.

You’ll no longer automate tasks – you’ll simply describe what you want. Yes, there’ll be some ridiculously highly-paid people who write the automation code that enables the environment to adjust itself on command, but if you’re not already doing that kind of coding, you’re not likely to be one of them. You’re about 5 years behind already, so you’ll probably just “consume” what they’re doing. But this changes how you manage your environment at a fundamental level – so start paying attention.


That is a hard word to say, “homogeneity.” Ask me to some time if you see me. Not at a bar.

Most corporate IT people lived in this fantasy world, a few years back, where we wanted everything to be the same. Same client OS, same server OS, same everything, as much as possible. We hated the exceptions – damn Marketing people and their Macs – and we did everything we could to contain them. We told ourselves it was cheaper to manage, cheaper to maintain, cheaper everything. We may have been right.

Doesn’t matter.

The decision was taken out of our hands when our users volunteered to pay for their own IT equipment. “OMFG,” the CFO said, “they want to do what? I’m in.” And welcome to the BYOD era. And it’s going to get a lot worse, because the diversification is going to get a lot more diversified. We can’t stop it – we already lost.

And yes, it’s a pain. And yes, it’s more expensive to support. And yes, it is in many ways less secure than an ideal world, but honestly your ACLs are a mess anyway, so let’s not drag security into the discussion. Anyway, you’re probably correct in every way that this level of diversification, and loss of control, is bad from a purely IT management perspective.

Doesn’t matter. We weren’t asked.

You basically have choices at this point. Choice A, bitch about it. Choice B, bitch about it and adapt anyway. Either way, you get to bitch about it – but Choice B gives you a career in 10 years, and Choice A makes you a relic. “Oh, this BYOD trend will stop,” you say? Heh. Go buy yourself a calculator: your organization is financially invested… in not being as financially invested in IT. Companies like Apple and Samsung are going to slowly work through the manageability problems, with able help from companies like Microsoft and Amazon, who can provide management interfaces. Remember, De Nile is a river in Egypt, not a career choice.

So what do ya do? Broaden the skill base. Learn to operate a Mac. And an iPhone. And an Android. I didn’t ask you to like it. If you liked it, it’d be a hobby, not a job. But you’re gonna need to know them all – and more, I promise – in a decade. Add Linux to the stack. And a few scripting languages; I suggest PowerShell, Python, and maybe Javascript. Diversify your portfolio.


I’ve been saying for eight years that Microsoft would one day ship Windows Server that didn’t have windows. Where it wasn’t even an option. And now we’re hearing early rumblings of “Nano Server,” and all I can say is… I told ya so. And if you think that this “Nano Server” won’t someday be the only version of Windows Server they ship… check back in eight years, and I’ll say it again.

So if you’re not already thinking like a Linux admin, with their keyboard-jockey attitude and funny haircuts, it’s time to change your haircut. Start thinking, “wow, how would I manage that machine if it didn’t have a video card anywhere in it?” Start thinking, “wow, how would I even get to that machine if it didn’t support direct attached human interface devices at all?” Yeah. Get to be an expert in networking again, and brush up on the remoting protocols, because these will have a big part in your future in a decade. Big, big. Trust me here. I’m only wrong when it comes to Transformers movies.


So with those in mind… what’re you thinking? Worried about anything? Think I missed a bit? Well maybe I’m just saving it for a future article, smarty pants. But maybe not – drop a comment and tell me what you think. Disagree completely? That’s cool – let me know. Be professional, and make a case for your opinion. Maybe, like Michael Bay has never done, you’ll change my mind.

5 thoughts on “IT Ops Career 2015: Where to Point Your Eyes

  1. FoxDeploy

    I agree, especially about infrastructure as code and declarative configuration management. As a consultant, practically every month I’m at a new client, and I see them suffering, they’ve got these delightfully hand-crafted servers, their little pride and joy machines with cute little names like the Three Magi or names of planets, things like that.

    And then as soon as we approach a project and need to build up a test-environment, its stumbling block after stumbling block to create parity in both worlds, to fully test things out.

    Sadly, some of the people in charge of infrastructure, particularly messaging systems, are the most entrenched against an infrastructure as code environment.

    I think it is resistance out of fear, mostly. If people could conceive of a world in which they could pop up three or four VMs, and then just apply their build configuration to these VMs and be ready to test, it’s just a whole world of pain, totally removed.

    My buddy spent all weekend at his shop over the weekend, rolling back a massive infrastructure change, all things which could have been managed using DSC…there I was, singing the praises of checking your infrastructure in to TFS or Mercurial, and just rolling back to the previous build.

    One of our greatest pains in the current world is the general lack of tooling in the Windows World for building configs off of existing systems and also the huge pain about stacking incremental configurations too.

    So, for the next foreseeable future, there will be a good living possible for those with the ability to explain these huge sea-changes to people with a willingness to learn.

    TLDR I agree.

  2. kayyum

    This is interesting and an eye opener.
    Thanks Don for sharing your vision on IT.

  3. David Jones

    So your saying that the market for creating DSC Recourse and Chef recipes it so saturated that it’s not worth the effort to get into it now? I would think that it’s still early enough that there is room for some more. Otherwise the last 8 months for me have been a waste.

    1. Don Jones

      DSC Resources? I wouldn’t call it saturated, and if you’ve been in it for 8 months, you’re probably one of the people I was talking about who’s *already doing it*. But if you’re *just now getting started* then you’re at the tail-end of the big opportunity. And if you’re *just getting into PowerShell,* then you’re not going to skill up to resources in time to make a significant difference in your career. The people who will make the big impact in that area already have their hands in the batter.

  4. Flynn

    Great read Don, I think its safe to say “Sys Admins” in 8 years will have a vastly different job and the gap between cutting edge IT organisations and say that of some Government departments will grow out even further as they refuse to update existing applications. Hopefully those admins stuck in the past don’t miss the incoming nano boat.

Comments are closed