IT administration is stressful?

I don’t find my job particularly stressful, most days.  The only time I experience significant stress is when my department is short-handed and we don’t have enough bodies to execute our projects and routine operational tasks.  Doesn’t happen too often, but when it does, we all feel the pressure.

In general though, IT is not a hotbed of hypertension and stomach ulcers.  So I was surprised to stumble across this gem, cited in a post on BlackBerry Cool:

The survey found that 36% of respondents indicated that their job is ‘stressful’ or ‘extremely stressful’. However, when email is down, 74% of respondents indicated that their job is ‘stressful’ or ‘extremely stressful’. Only air traffic controllers and police officers were viewed as having jobs more stressful than that of an IT administator whose email system is down!

With respect, if you’re that stressed out, something is very, very wrong.

The BlackBerry Cool post goes on to enumerate five steps to make BlackBerry server administration less of a hassle, but which also apply broadly to IT endeavours in general.  I’ll summarise them here:

  1. Ensure adequate capacity
  2. Minimise network latency
  3. Monitor and audit your infrastructure
  4. Judiciously apply patches (be aware of potential impact to other systems)
  5. Monitor end-to-end quality of service, internally and (where able) externally.

Sensible stuff, but the best stress-relievers are not included at all.  My approach to IT stress reduction can be boiled down to two main points:

  1. Know your role
  2. Build appropriately

KNOW YOUR ROLE

Most IT people tend to focus narrowly on the technical aspects of their jobs and do not really take the time to seek out personal networks and interaction with the business side.  I would characterise most IT jobs as equal parts salesman, civil engineer, detective and doctor.

Salesman?  Yes, IT guys need to be able, sincere communicators.  Most of them aren’t, and this I would say is the root of the problem.  Our job is to support the enterprise, and sometimes the business side doesn’t always act in its own best interests.  Your job is to sell them workable, affordable ideas and solutions.  Most importantly, you have to convince the business that every penny spent in IT pays greater dividends to the rest of the company.  If you can’t do that, you are going to make your job ten times more difficult than it needs to be.  You won’t get the resources that you need, and the business will not get the level of service it needs.

The civil engineering aspect is obvious.  We provide solutions that integrate the physical and natural worlds, that allow flesh and electrons to interact in a way comprehensible to both.  We must also design for evolution of the infrastructure, to allow today’s foundation to host tomorrow’s innovations.

We need to be able to investigate, isolate and prevent certain activities within the infrastructure.  And we need to be able to rapidly identify and treat a wide range of conditions that afflict single users or an entire enterprise.

BUILD APPROPRIATELY

The BlackBerry Cool post mentions that stress happens when systems are down, and obviously so.  Having a downed system is, however, an entirely preventable situation.  Mission-critical systems should always be built for fault tolerance and redundancy.  Always—there are no exceptions.  Today’s servers are guaranteed to go offline, whether due to routine maintenance or unexpected catastrophe.  It is not optional or unexpected; these are human-designed systems with undiscovered faults built-in.  The bottom line is, if the business is not willing to fund a mission-critical system with fault tolerance, then that system is clearly not mission-critical.

At The Firm, all of our application and mail servers are clustered.  This means that two or more servers have copies of the same information, and any changes to the data are immediately replicated to the other servers in the cluster.  Yes, it means your hardware and network costs are more than that of non-redundant systems.  On the other hand, if a server dies, it’s no big deal.  Traffic is automatically routed to the failover servers, and the change is invisible to the end-users.  Wave goodbye to any pressure from the brass to get the failed system back online.

Building redundancy into your mission-critical systems also means maintenance at your leisure.  You don’t always have to wait until after business hours to perform maintenance; you can take down a piece of the infrastructure at your leisure.  Again, the end-user is not inconvenienced, and neither are the IT staff.

This may mean wringing additional funding out of the company in order to have clustered mail or application servers.  You don’t have to go whole hog and cluster everything, though—just the things that the business can’t run without.  Does it cost more?  Absolutely—this is where your salesmanship comes into play.  What do you gain?  An end to outage-related stress, for both IT and the business.  Isn’t that worth a few extra bucks in the budget?

Category: Industria, Web/Tech
You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

Comments are closed.