by Peter Wayner

6 reasons to invest in enterprise architecture tools

Opinion
08 Apr 2021
Enterprise Architecture

EA software can bring order to the innerworkings of your enterprise’s sprawling technical operations. Here is the upside of mapping and cataloging — and some cautions about where the real magic resides.

Enterprise architecture is designed to bring order to the vast realm of corporate IT and its burgeoning collection of machines and software, a cornucopia unimaginable just a few decades ago. Desktops, tablets, phones — screens as far as the eye can see. And that’s not counting screenless gadgets like sensors, audio devices that answer to “Alexa,” and everything else from the so-called internet of things.

Alas, the digital abundance isn’t always a gift. After all, it all comes with responsibilities. The machines, algorithms, software upgrades, and the colo farms seem to multiply and someone needs to watch over them and care for them.

At their core, enterprise architecture (EA) tools maintain a master list of assets in the form of a database table that catalogs the devices. Then it adds a collection of features for displaying this information in a way that’s easy to use. The magic comes from the team that deploys the software, fills it with data, and then uses it to make smarter decisions.

Is all of this tabulating, curating, and tracking worth it? Is the fuss worth the bother? Here are six reasons why investing in an enterprise architecture solution can be worthwhile for your organization — and some gotchas to keep in mind when relying on EA tools.

Order is good

Anything is better than a vast network of uncounted computers running unknown software written by untracked workers who may or may not work for the company anymore. A list written on paper would be a start. A spreadsheet would be better. Enterprise architecture tools are light years beyond a list. They add order to the world, providing a wealth of information about the vast sea of bits sluicing through your enterprise’s endless collection of hardware.

It’s important to remember, however, that tools do not provide order; people do. Enterprise architecture tools just provide the means for establishing order. Imagine, for instance, a note saying not to turn off a server without talking to the customer acquisitions team. What if that team was dispersed three reorgs ago and no one updated that note?

Software developers have often discovered that documentation can cause problems when it diverges from the code. Suddenly people are assuming one thing while the code is doing another. Enterprise architecture tools are still the solution, but they aren’t magic. They come with a commitment to maintaining the data. They’re just a path for your team to bring order. They don’t bring order on their own.

Enterprise architecture breaks down silos

Organizations can suffer from isolation as differences multiply. One team chooses technology A while the other chooses B. Installing enterprise architecture software won’t fix these deep differences, but it will make it easier to spot these differences. The process of cataloging enterprise assets in an enterprise architecture tool reveals many of the distinctions and that is the first step to building some unity. The central database is a catalyst for change.

Of course, just because inconsistencies can be surfaced by EA tools, doesn’t mean they should all be resolved. We may pull out our hair condemning the fools that cobbled together the collection of barely communicating software packages, but these differences often offer a little security and resilience. Monocultures are easier to attack and are more vulnerable to catastrophe. A problem that affects one corner will affect all of them. This isn’t to say the mental cacophony of a disharmonious software collection is pleasant or desirable. Only that it has some advantages.

Metrics identify problems

The lesson from the business schools is clear: You can’t manage what you can’t measure. EA software brings a collection of tools for measuring how the digital realm is functioning. Now there is a way to compare teams, departments, and divisions with each other. EA makes it possible to spot the laggard servers, the overloaded databases, and the overburdened networks.

Remember, though, that metrics can bring noise. The data is out there. It can be aggregated, cleaned up, and displayed on a slick dashboard. But just because the bits are available doesn’t mean that they bring any real insight. The tools offer a wealth of information about your enterprise’s assets, but it takes expertise to know just what the numbers mean.

Automation saves time

Many of the EA tools are justly proud of their ability to connect to the computer networks and gather quite a bit of information automatically. We’ve reached a level of development where there’s already quite a bit of telemetry available through APIs and debugging tools. This can all be swept up into a useful interface so we can see what’s going on.

Of course, automation is rarely bulletproof. The plot of Michael Crichton’s novel Jurassic Park hinges on the failure of a computer program to take an accurate census of the dinosaurs on the island. It’s only programmed to track the dinosaurs that it knows about and so it misses the fact that the dinosaurs are breeding. This blindness is also a danger for EA systems. They can only follow the computers that are integrated into it.

A single source of truth is essential

Too much time is spent searching for the correct information. Enterprise architecture tools act as a single source of information that saves time. It can be expensive to build out the infrastructure but once it’s running, the truth is easy to find.

At the same time, it’s important to know that, despite EA software’s promise of providing a single source of truth, single sources of truth rarely exist. Even when they do, they can be wrong and there would be no way to know. When there’s only one opinion, there’s no consensus and no opportunity for analysis.

Still, a flashlight, no matter how weak, is better than fumbling around in the dark. Rather than obsess about how the software won’t likely perfectly map your enterprise, it’s better to put aside these hesitations and get on with the difficult task of cataloging and tracking the cornucopia of data processing your enterprise has delivered.