Avoiding SysAdmin Obsolescence; The Only Thing We Have to Fear is Telepathic Robot Drones

(Originally posted at my blog here, but I thought I'd share it on the SysAdmin-Network as well to get a more targeted response. This is a topic of great importance to me in the last few months. Flame on!)

A recent article at queue.acm.org piqued my interest. Titled “Overspecialization can be the kiss of death for sysadmins”, a budding SysAdmin asked “What is the biggest threat to systems administrators?”, specifically referring to the SysAdmin profession.

The answer was a 14 paragraph monologue that, in my interpretation, was very good advice but answered a completely different question. The answer boiled down to two points:

1. Work smarter, not harder
2. Communicate better with your bosses concerning what it is that you do all day

I think that is great advice for keeping your job today, but doesn’t seem to answer the bigger question that the budding SysAdmin asked concerning where the SysAdmin profession is going as a whole. Either that or I’m hyper-sensitive to this topic since I’ve been all abuzz about it in the last few months and I’m misinterpreting everything I read.

First, I’ll address the well meaning advice that was offered to the budding SysAdmin.

Point number one of the original article states that it should be a SysAdmin’s priority to strive to know what to do more than merely how to do something. You need to be a decision maker. Those that are merely grunt workers, churning out mod_rewrite rules on command or twiddling ACLs in the ASA are easily outsourced. Those that primarily decide why a rewrite rule is needful or what threats need to be mitigated at the gateway are the valuable ones. Double points if you do both. In fact, it is my opinion that the separation between decision makers and implementers is closing. due to cost and a greater potential for miscommunication.

The article also lumps overspecialization into it’s first point. Not only is it bad to stay comfortable with being told what to do rather than learning why to do something, it’s equally bad if you only know one thing very well. If demand for that thing goes away (and it always will, in some capacity), you’re left out in the cold. However, in my opinion, that’s merely a symptom of choosing to think about how to do something rather than why you should do something. It’s one thing to know everything about how to run a web farm using NCSA HTTPd. It’s a more valuable thing to know why you shouldn’t. Overspecializing is a symptom of not thinking hard enough.

The number two point in the article points out that even if you’re awesome at what you do, make great decisions and implement stunning examples of systems engineering perfection, if no one knows about it then it’s essentially useless to you and your department. That’s true enough. People need to know what you do so you can justify your budgets and positions. This isn’t about slimy self-preservation and self-promotion. This is professionalism. If you want your company to succeed, then they need to know what you do, why it should be done and how much it costs. If you communicate well with those around you, you will have less frustrations and more success. Fairly simple, but it’s a tried-and-true concept.

My argument is that this has nothing to do with the original question about the SysAdmin profession. Any professional in any field should be heeding this advice, which is great, but isn’t addressing the larger shifts in the profession towards outsourced and decentralized IT that are being seen.

A budding (or veteran) SysAdmin will probably be confused and angry if he defines his role well and communicates with his superiors, peers and users and yet still finds that his job at an internal IT department is hanging in the balance or changing into something that’s terribly foreign to him.

Well defined job roles and good communication are essential, but we’ll probably have to be doing this from within individual business units or from outside service providers.

I will give my own answer to the budding SysAdmin that asked a very good question: “What is the biggest threat to systems administrators […] as a profession?”

The biggest threat to Systems Administration as a profession is: Telepathic robot drones.

Seriously, until machines become sentient and can read the minds of the directors and executives of a company, I don’t think we’ll be obsolete. Our job is secure. However, don’t expect our roles to look the same in 10 years. Not even 5 years. I think there will be fewer specialized techs within corporate IT departments and more of them will cluster into service companies that provide their specialty with great precision.

I do not believe that all kinds of specialization are a fast track to obsolescence. If you focus on one solid thing (operative word being solid; COBOL is not solid), such as storage area networks, network security, Active Directory, you can develop along with the technology. You can watch trends and see the signs of change better as a specialist as a generalist. From there, it’s a simple matter of choice to change with the times before you’re left in the dust.

Internal IT departments will become more mediators between the outsourced specialists and the needs of individual companies or business units. Internal IT people will take care of the company’s fragile, custom solutions or systems that are not trusted to outside providers such as databases with sensitive information.

In the end, as long as there are computers there will be SysAdmins. We may become a bit specialized and we might not have the same opportunities to work internally to a company, but we’re here to stay.

Oh, and the advice about working smart and communicating? That’s true for any profession whether you’re a plumber, welder or pogo stick designer.

And you? Are you specialized and worrying about obsolescence? Are you a generalist and feeling secure in your position? What do you believe about the future of systems administration?