Workshop Review: The State of the Profession

I was scheduled to attend a tutorial today, but I pulled a last minute audible so that I could jump into a workshop on a topic that I spend a lot of time thinking about: The State of the Profession: What Are the Unresolved Issues in System Administration

The workshop was organized by Tom Limoncelli and Kyrre Begnum, both well-known administrators with a long history of attendance at LISA. The workshop was very well attended, with between twenty-five and thirty administrators who came to discuss what they felt were the issues which should have been solved, but for various reasons, weren't.

One of the more interesting features to me was the wide ranging scope of the topics, from software packaging and OS integration to the definition of the profession itself. Tom and Kyrre will be submitting a summarization of the workshop to ;login: for later publication. While there were no resolutions, I think all of us came away with a greater understanding of each other's concerns and duties. Here are some major topics of discussion which I hadn't considered fully before:

• A lack of perception of cost of IT services
We've all heard, "why is enterprise storage so expensive when I can buy multiple terabytes for under two hundred dollars?!?", but another consideration in the cost equation is the expense of time. I have spent so much time arguing monetary budget that I don't think I ever really tried attacking the "cost per hour spent" metric. I've considered that free is usually free in name only, but it goes much deeper than that, particularly when the size of the team scales up.
• Where are the PhD SysAdmins?
An argument can (and has) been made that ours is a practicing profession, not an academic one, but given the ever-increasing number of system administration programs around the world, this isn't likely to stay the case, but there is still a conspicuous lack of doctorate-level administrators and architects.
• We are not recruiting the right people to become administrators
This thought had never even crossed my mind before the gentleman in this session brought it up. I, like most people who are currently administrators, was trained "on the job", so to speak, under existing administrators. In university programs who create system administrators, there appears to be a trend of attracting people who "filter down" to the job - people who don't want to program, people who don't want to become superstars, and people who aren't outgoing. Yes, we can absolutely make them into good, solid administrators, but his view was that they turn into granite, whereas we need some diamonds. I'm not necessarily sure that I agree, but it is interesting that we apparently act as a "catch all". The gentleman presenting this topic believed that it was because we have a relatively low degree of prestige. We don't make the kind of money that neurologists make, we aren't given levels of autonomy that high-level programmers are given, and we're unlikely to become world-famous at what we do. We don't currently have, in his opinion, the draw that brings great people to us.

There were more, but to get a better idea, you should read the article in ;login: when it comes out. To get ;login:, you need to be a member of USENIX, so join here.