Fast forward 10 years. I'm now in Silicon Valley, and the company I am working for has made a serious move towards Unix. They are also trying to get SAP implemented.
Well, I couldn't do anything about SAP, but as one of the Unix people, and as a manager of a systems support group, I'm asked by the IT group to come to the computing center and help train the phone people working there on Unix. This is not rocket science--most of the phone people are local university students, and I've taught an Intro to Unix class in industry and at UCB extension. So, one sunny morning, I take off to the computing center, about 70 miles up the road.
The training was pretty uneventful, everyone behaved themselves, and I provided them with a "20 most likely asked questions and the standard answers" sheet. I had an hour before I needed to head back, and the shift manager at the computing center decided to give me a tour of the operations.
The machine room looked like a mini-mainframe swap meet--there were Amdahls, Tandem proprietary boxes, IBM VM and MVS systems, some random Vaxen, and one lonely Sequent mega-box. In all they were running nine different databases on eleven HW/SW combos. None of the databases talked to any of the others.
I commented that once the SAP systems, running on AIX servers, were brought in, their 5000 sqft machine room was going to look a little empty.
The shift manager gave me a look. She said: "Look around the control room--how many operators do you see? We employ at least three people; an operator, a sys admin, and a DBA, for every box, for every shift. We're been told that when SAP comes in, they'll lay off 80% of us. Do you think we're really interested in that? They'll never install SAP here."
And SAP wasn't installed there. They managed to find all sorts of interesting ways to put off and break the installation from the HW on up. Two years later, Europe and Asia had SAP running, but North America didn't. The VP of IT was let go.
Only when the corporation had a CEO who took a good hard look and canned 80% of corporate IT for being more interested in keeping the status quo than getting the coroporation into the 1990's, did SAP get installed.
A few morals here:
The VP of IT should have never told the group that once the magic system was installed, he could get rid of 80% of them. Turns out that he was trying to win the "good dog" award by getting way under budget.
Even if you consolidate on one platform and one system, you will likely need more than one set of people to take care of things. After all, you're still going to be crunching a good proportion of the the current job load-- or more, since once systems are consolidate, groups tend to ask for new stuff.
If you're an IT person who's made their mark with a now aged system, you'd better expect that someone is going to want to upgrade it. Your job is to make yourself useful enough that you're brought on board to learn the new system.
There are people who manage to drag things out to the bitter end--I worked with one CS person at university who knew the insides of NOS (old CDC 6-bit byte operating system) like no one else. After the university had no more use for him, and he refused to learn Unix, he ended up at the old ihnp4 (Indian Hills Naperville AT&T site), tending their antiques. Not a lot of future in that, though.
If you hang on to an old system, you can expect things to end badly.
Note: This is different from an unpopular or politically incorrect system, although the outcome might be the same.
And finally, the CEO owes it to their corporation and their shareholders to pay attention to what's going on in IT. If the IT organization goes through Directors and VPs like Chiclets, there's a big mismatch going on, and it's not a good thing. Don't expect that your IT organization will run on auto-pilot. It always amazes me that while a CEO will pay close attention to engineering, the part of the company that books and bills gets relegated to the garret.
Posted by lsefton at September 21, 2003 08:56 PM