h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

I've been writing a series of journals about best practice in IT departments. Over the last couple of days I've described what is the difference between what we would call an 'incident' and a 'problem', and how we manage them.

Both of these things occur day-to-day, of course. We have permanent teams of people on help desks and in teams of engineers who deal with incidents and problems as they happen. This is what we call 'operational' work. Some people in IT call it 'business as usual'. Everything we do in response to these is designed to get the system up and running again as quickly as possible, then making it stable and working as it was designed to do.

There are many other things that IT people do - involved with designing, building and implementing new systems - and I may talk about some of these in due course. Before I do, I wanted to mention a couple of other things that the help desk and the operations teams do from day to day - they're not just larking around in the office waiting for the phone to ring, you know! Well, unless you know different, that is.

There are three other operational processes that people follow to keep your IT running.

One of them is to monitor your systems to make sure they are running correctly. Technicians call this 'housekeeping', or maybe 'daily checks'. Basically, one technician - maybe one from each team - gets into the office early and runs through a series of tests to make sure things are working fine before the start of business. They will look at logs and system reports. They will check networks and servers and databases and storage systems. They'll also log in to application systems.

We also use automated monitoring systems - software which pings your devices, checks available memory or storage space or processor utilisation, and simulates logins. This is a process which we call Event Management. Typically, these monitoring systems can produce hundreds of thousands of alerting messages to say what they've detected. It's quite an art tuning the systems to react properly if they detect something serious - maybe they would send an e-mail or an SMS, or make a red light blink on a dashboard in an operations centre. Clearly, you don't want false alarms, otherwise people might ignore the light even when the event was serious.


One other process that a help desk does is to handle callers' requests as well as their incidents. Requests tend to be simple defined tasks that a help desk agent can perform without needing an engineer to help. Simple ordering tasks or password resets are good examples. They will also take your complaints as requests - rather than arguing with you, they should politely listen, take the details and pass it on to the right person to deal with.


Finally, there's a special type of request which the help desk will often receive - granting people access to applications. Maybe somebody new has joined the company and needs access to the e-mail or HR or finance systems. Similarly, when people leave, their accounts must be removed for security reasons. The ITIL books call this Access Management. Generally, the help desk has to contact managers for approval, then contact systems engineers to set up the required logins, so this can often take a while to complete.

[Amy P]

Since I know zilch about IT, pretty much everything I now know about it has come from Icy's journals.

Thanks paul, I appreciate it's not always a very interesting topic, so thanks for reading!

It's really crucial to business that Help Desk is full of people who understand and resolve issues. I love when it works well.
Thanks!

it is interesting