3 Key Things Many Data Integration Strategies Overlook

Data Integration So, you’ve started down the path to data integration.  You have a strategy, an approach, and you’ve even picked your best of breed technologies.  The bad news?  You might have missed a few things.  Worse, you won’t know about them until much later, when everything is more difficult and expensive to fix.

I’ve identified the three most common ‘misses’ in a data integration strategy.  These are omissions that lead to real problems down the road, and they are easily avoidable.

First, a lack of a systemic data governance approach.

Data governance is not just for database systems; it should be systemic to all data that’s persisted, and the data integration approach and tool needs to support data governance.

In other words, you need to have the ability to place policies that can determine data’s use, ownership, security, compliance, and anything else that may be relevant to the data element.  The idea is to protect your data from yourself, and insure that, as it moves from place-to-place, the policies are consistently enforced.

Second, lack of a performance model and plan

The number one complaint that I get when asked how the data integration project is going is that performance falls short of user expectations.

These are not technology issues; they are planning and modeling issues.  Meaning, planning and modeling was not properly done, and thus the data integration solution is not optimized for the use cases.

When added to a data integration strategy, the issue of performance is dealt with up front.  That means no surprises, when it comes to how well the system meets performance expectations, considering that it’s all been modeled, tested, and documented.  Moreover, most performance issues are traceable to design and deployment issues, such as not doing complex schema transformations in the most effective and efficient ways.

Finally, lack of data integration security, and links with enterprise security

The types of security services that you need when deploying a data integration solution vary a lot, depending upon the requirements of the applications and the data.  As a rule of thumb, you need to deal with identity and access management, encryption at rest, and encryption in flight (data moving from store to store).

Often neglected are the links of your data security solution at the data integration level, and the security solution for the entire enterprise.  They have to work and play well together.  If they don’t, they’re really working against each other.  So, make sure to be inclusive with the enterprise security team, and have ongoing proactive communications.

Data integration is a core part of any successful enterprise IT program.  Indeed, it’s often the most important way to spend your budget.  However, like anything worthwhile, you need to have a strong plan behind it, and a strategy that really hits all of the major issues.

Comments