Letting go: the hardest part of some careers

You already know you’re among the “data center personnel [who] need to stay on top of what skills are hot …”, as Paul Korzeniowski phrases it this week in “IT pros adapt to stay relevant–and employed–in IT jobs field“. It’s more than just keeping up and readiness to learn new material, though. To some extent, changes in employment dictate what you must give up.

IT skills don’t just cycle economically the way other broad categories do–one year, public-school teaching is competitive, the next, it’s nursing and transportation, and eventually it comes back to teaching. Several of the skills or specialties Korzeniowski rightly characterizes as “not hot” aren’t just out of favor temporarily. They’ll never return.

That realization results in a different attitude about careers. Not only must you be eager and able to learn new material, but you also need to recognize what to leave behind. To save the results of a runaway sixty-hour computation by cleverly attaching to an unused hardware port with a spare VMS workstation and using cat and dd to get at the process table directly might have been a high point of your career. Now, such an exploit belongs in the “hobby” part of your resume, because it’s cheaper to buy a new host, or even cheaper to rent a few minutes on a supercomputer-in-the-cloud, than your time to be clever this way is worth. Down-to-the-metal, syscall-level knowledge still is demanding and difficult and occasionally effective; I will admire your expertise in it. At the commercial level, though, demand for work of this sort is confined to a few very specialized operations, and won’t ever grow back.

There’s no particular value in arguing that organizations don’t realize what they’re giving up with this, because the gains from more abstract or higher-level management of inexpensive, replaceable commodity computing elements more than make up for the loss of individual, low-level attention to assets.

Also, be realistic about where your career targets fit with corporate plans. An old-style datacenter might have been staffed with skilled sysads intimately knowledgeable about every host and network element. When an organization constructs a new datacenter, it doesn’t plan to replace all the staff with people who know the hot new skills of virtualization, software-defined networking (SDN), and so on; the organization expects most of its on-site human assets will be more like $15 / hour operators, guards, and watchmen. That’s the future for anyone who chooses to stay exclusively on the datacenter front lines.

If you’re open to a little adjustment, though, good opportunities remain. Much of the adjustment is in attitude. CFEngine founder Mark Burgess describes the role of his product line in these words: “We want to help sys admins transform their traditional role from being reactive to agile and proactive–to become valued infrastructure engineers!” That transition might be bittersweet for someone who excels at “fire-fighting”, of course.

You can, however, make the most of your diagnostic insight and experience by shifting their focus slightly to matters of system design and business management. The tenacity and alertness that have brought you this far apply there, too. Learn hot new topics–SDN, RFID, security certification, mobile device management (MDM), and more–along with business skills and perspectives, and you’ll be fine. The only realistic alternatives are to confine yourself to the shrinking collection of organizations still using old-style hands-on operations approaches, or to accept a salary that roughly equates you with a pizza deliverer.


  1. Vasudev Ram says:

    Good post.

    Apart from the overall content, which is good, I liked the point about fire-fighting. I’ve seen that happen a lot, over the years, in system administration as well as in software development and management (having worked some in all three areas. A common statement in more than one place I worked at, was “There’s a fire”. But not many people, senior or junior, thought of preventing those fires instead of fixing them). Some proactive measures up front can mitigate the need for fire-fighting a lot.

    – Vasudev

    • cameron says:

      Maybe there’ll be time in the next few weeks to follow-up with an outline on how to take a data center from fire-fighting to “pro-activity” …

Leave a Reply

Your email address will not be published. Required fields are marked *