Data Integration Blog -

We’ll it’s after the New Year, but January is not a bad time to list things that IT can do to improve data integration within the enterprise.  In many cases it’s time to approach data integration for the first time.  In others, it’s a matter of tuning things up.  Either case, it’s a good idea to access the existing capabilities, and seek improvements from time-to-time.

So, here are the top five data integration New Year’s resolutions you should consider at the beginning of 2012:

1.) I resolve to define or refine my data integration strategy.

What?  You have data integration technology in place, but no overreaching strategy that includes a data integration technology implementation roadmap?  You’re not alone, but it’s not too late to define a new strategy, or refine the one you have.

This is a matter of a ground up assessment of your existing systems and data stores.  You need to define the value of data integration technology that’s able to drive an information exchange between one or more data stores.  I would typically start by defining the value of each integration instance, then put them in order by priority.  Next, begin the implementation of the data integration technology…moving from problem domain to problem domain…from the most valuable to the least valuable.

Writing this down makes things much easier to budget and plan for.  This will also ensure you have buy-in at all levels.

Moreover, data integration is an aspect of other strategic IT transformation projects, such as moving to better-governed enterprise architecture, or a SOA.  You need to make certain your strategy meshes with those efforts as well.

2.) I resolve to define my data semantics and inter-relationships of data.

This is missing in almost any enterprise that I’ve worked with over the years: A common understanding of all business data managed within the databases, be they old or new.  There are typically many ways that the concept of a ‘customer’ is defined, as well as ‘inventory,’ a ‘sale,’ you get the idea.  Which way should be the common way?

We need to figure out what the data means, and where it exists.  Moreover, what is the data of record?  Once we have this better understood, data integration becomes much easier to define and leverage.

3.) I resolve to make sure that data integration is a part of any cloud computing migration project in my enterprise.

The good news is that we’re moving to the cloud.  The bad news is that little thought has gone into how data integration exists within the target cloud platform.  At the end of the project we could have just another silo, but it’s in the cloud this time.

Data integration is almost always a requirement of any migration to cloud-based systems, be they IaaS, SaaS, or PaaS.  However, it’s often an afterthought or forgotten all together.

I spend a great deal of my time defining data integration approaches around the use of cloud-based technology, private and/or public.  The critical success factor there is to make sure it’s systemic to the migration planning, and the solution is in place by the time the cloud-based systems move to production.

4.) I resolve to check my data integration technology.

In many cases, data integration technology is several generations behind the times.  What was a good solution in 1999 is probably something that should be updated in 2012.

This is not to say that we’re forcing you to spend money on technology that you don’t need.  However, the process of looking for better solutions to fit your data integration needs may actually save you money when you consider the value of the new features that new technology may offer.

I would call this a technology refresh, back in the day.  This is a time when those charged with operating core systems – in this case, data integration systems – determine the value of updating or replacing existing data integration technology solutions.  Many of them are more than 10 years old.

This is a bit scary.  There are inevitable problems that will need solving.  However, the value of this exercise should produce much more value than problems.

5.) I resolve to understand the architectural value of data integration.

As I introduced above, data integration is systemic to any good enterprise architecture or IT strategy.  However, many skip, forget, or ignore it, and have to back in tactical solutions later in the game that are not well understood in the context of the architecture.

Now is the time to understand the value of this technology, and why we need to pay attention to it.  Typically you’ll get as much as 10 dollars back for every dollar spent on data integration, people and technology.  Few things in enterprise IT these days provide that kind of ROI.

This means accessing what data integration is to the modern enterprise.  Moreover, you need to access strategic technology that allows systems to freely share information, thus adding efficiencies and effectiveness to core business systems.

Good luck in 2012.