Executive Advisory Program: Real Legacy Issues

Posted by Chris Howard

My apologies. It has been a long time since I posted. Jack has been valiantly holding down the blogging fort while I've been on the road. Mike Rollings was off the grid for a full two weeks, resting his brain, seeing his kids through graduations, and tending his new scruffy beard. Well, I'm not so sure about the last part...but beards are back in style (i.e., Joaquin Phoenix, Mike Weir, Chris Howard).The last three weeks, I have seen nine countries and as many cities, ranging from Basel to Phoenix.This is my first week in my own office chair since the end of May. And my brain has settled enough to write a post.

While in Rome during my hiatus, I was faced with an onslaught of history and metaphors. After a few days there, one gets desensitized to the oldness of the place: "Oh yeah, there's another ancient Roman monument." You eat pizza in front of the Pantheon nonchalantly. You rub your tired calves on ancient paving stones. You smile brazenly for pics in front of the gaping guts of the Colosseum. History is a fact of life in a city like Rome. You take it for granted.

Until you have to build a new subway line.

There is a project in Rome to construct Metro Line C. The line cuts across the city along the line of popular tourist locations. As you walk the narrow streets and alleys of the Ghetto and Campo di Fiori, there are signs of work below proudly advertised with official Metro signage. And more often than not, signs indicating archeological work.

Rome has rebuilt itself by building on top of itself. Layers of civilization are everywhere, most of it unexcavated. In the centro storico, dark stain lines high on the walls indicate where street level was before excavation began. As a result, if you dig a hole, you need to be prepared to work closely with archeologists charged with preserving Rome's legacy. If you are tunneling under the city to create a new subway line, the legacy problem is orders of magnitude greater. Project plans are being adjusted as we speak to extend the "completion date". In fact, one sign I saw gave the project timing as "Spring 2007 until Completion." How's that for non-committal? As one Metro official said, "We expect construction of Line C to take between five and seven years. It is impossible to make a more precise estimate because so much will depend upon the archaeological finds." That was 8 years ago. Completion is now targeted for 2015.

So, techno-geek that I am, I immediately draw a comparison between this scenario and the challenge of legacy IT infrastructure and applications. When I explained this to my 18-year old son who was along for the trip, he didn't find it very exciting. His response was "why do you think about this stuff?" But I digress.

The comparisons that strike me:

1. Some of the historic substructure is mapped and known. Much of it, however, is not. You don't find it until you start to dig.

2. Once you hit something in the legacy infrastructure, it's unclear how deep it goes, how far it reaches, and what dependencies it contains.

3. It is probably unacceptable to just dig out and remove legacy stuff. Worst case, it is supporting something else that comes crashing down as a result. More often, it has tremendous meaning/value to someone who therefore blocks its destruction.

4. The best intentions of providing something new and highly efficient are subject to delay as legacy issues are resolved.

5. Access to the new system may end up being convoluted due to complex existing, immobile structures. (In the case of Metro C, this means carefully considered station access points, and there isn't much space to work with.)

6. The resulting solution will not be a straight line. It will bend around legacy structures.

7. Some projects may be abandoned because the risk posed to the superstructure is too great.

So, the moral is that our IT shops are quite Roman. We have extended our IT infrastructures by building on top of older versions. When it comes time to put something new in, we are faced with a fuzzy set of legacy dependencies that increase risk, extend project timelines and complicate design. Maybe we need a better city model. Phoenix? Not so much.