Tuesday, 19 October 2010

Enterprise Landscaping

Behold the second piece of my garden, together with some part of the bridge. The pile of tiles is there to nag me so I'll leave Twitter and blogging alone and get to redo this piece of garden

Redo? Yes, redo. When looking for inspiration today, Henk van Zuilekom was kind enough to give me "one of those tasks":
@ draw a parallel between re-doing legacy applications and re-doing a garden. where does it all go wrong?
I wonder whether I inspired him as I posted the picture above a few days ago, but I do see some very useful parallels between redoing enterprise IT landscapes, and "just landscaping" your own garden

Coincidentally, tonight's tweets and blog posts from other have been pointing in the direction of open hunting season on IT architects, somehow - so maybe it's a good time to reintroduce common sense, which this post is all about

How do legacy applications get redone? There is one story to that, with a few chapters, none of which always get read. Usually, the bottom-up approach is wielded by IT:
  1. Look on all hard drives for applications
  2. Make a list of all applications on all hard drives
  3. Hand over list with applications to IT manager
That is very much infrastructure-like: start at the physical AS-IS, in order to achieve the logical / conceptual TO-BE. Works, according to The Book. No, don't have any copies of that, just in case you were wondering...

Here's the software architect approach:
  1. Make a list of all application managers
  2. Interview them
  3. Hand over list with applications to project lead
That's very application-like: check, then double check. It's driven by a background in coding where review is a mere formality, sometimes even felt as a lack of trust, even.

The Architect approach is to ask the users, what they are using, and then cross-reference that with the backend. Still, the ultimate end-goal is the list of applications as mentioned above.

The result? A messy mismatch between that what's there, and that's what supposedly in use - and then what?

Redoing a garden is a whole different story. There, the owner points out what he wants to keep and what he doesn't, and how he would like it to be. The gardener comes in after of without the landscaping architect, plants and builds everything according to the rough sketch, money passes hands and the customer's satisfied

Wow. That's quick. In IT, we had only a list that didn't add up much. Where did it go wrong?
  1. Overseeing a garden, even a big one, is a lot easier than overseeing an enterprise IT landscape. The difference is that people are actually using IT tools that aren't in the overview. However, much like people are making apple juice and marmelade off of a garden, those should just be considered out of scope
  2. Ownership in a garden is much less complicated, and a lot better 'documented'. There never is a tree or plant unaccounted for: if there are, they're just considered weeds, and exterminated
  3. Functionality of plants and trees is either blatantly obvious, explained by the owner, or irrelevant. No need to seek high and low for hidden features, advantages, disadvantages or risks
  4. The so-called AS-IS never is very interesting in gardening, only the available space is. People have an idea about the TO-BE situation: they look ahead, not behind. Sure, a special plant or very old tree might survive, but there aren't any "action groups" advocating the survival of some ugly plants or trees unless - well, unless the garden gets to be a few dozen square miles, and public property...
A simple garden is just a garden - but when its size approximates an enterprise IT landscape, it will be just as complicated as one. Remember protests and fights over a forest or nature reserve threatened to be leveled for a highway or train? Having to dig through dozens of archives to retrieve ownership of a land? Interviewing whole towns to be able to value a piece of property?

Ah, now it starts to make sense, doesn't it? Current enterprises are just too big, and the supporting IT landscapes dito. The unknown that is considered weed in a simple garden, is a potential huge business risk: what if entire countries depend on the apple juice and marmelade our garden trees and plants supply?

The solution is simple, especially now: chop up the enterprise, redo that in stead of only the IT department

The enterprise has become too big too sustain, from a human point of view, and now also from an IT point of view

2 reacties:

Opie said...

Hi Martijn, thanks for the mention and, yes I was triggered by your photo and that tremendeous task of creating that pile of stones a few days before that.

One thing you missed is that when redoing a garden, small or big, is that no-one really cares whether the garden is usable in the mean time. Bring in the machinery and dig down a couple of inches, uprout everthing, rip it all out and basically: create a greenfield. Try that in a big organisation and you're out before you can finish your coffee.

But with gardens, things need to blend as well, thats the difference between the amateur garden and the well-designed, balanced garden. You can try growing those colory things (flours or flowers or something like that) under those trees, but it will not be a succes. Much like some applications that get created without properly thinking about the (process-) environment in which they need to bloom.

In both cases though, you would like to look at what you want. Not hindered by anything you already have. Then, and only then, should you look at what you have and what you can salvage to fit into your new landscape. In gardens that is the most likely approach that will be taken. In IT, its usually: lets see what we can salvage and then fill in some gaps.

Don't forget a place for the barbeque.


Martijn Linssen said...

Thanks Henk! You are absolutely right, this piece of garden has been out of use for a few months now, and no one complained - then again it's only 7% of my total garden; maybe something to keep in mind at the next enterprise reshuffle

But yes, we should look ahead first, not back - and if users tell us they just want to keep using what they're already using, we should just ask them what it is, and not swap architecture for archeology and try to find out for ourselves

No worries about the BBQ, this wil become 35m2 of tiles! No more weeding either

Post a Comment

Thank you for sharing your thoughts! Copy your comment before signing in...