It’s time to consider data-driven enterprise architecture.
The traditional approach to enterprise architecture – the analysis, design, planning and implementation of IT capabilities for the successful execution of enterprise strategy – seems to be missing something … data.
I’m not saying that enterprise architects only worry about business structure and high-level processes without regard for business needs, information requirements, data processes, and technology changes necessary to execute strategy.
But I am saying that enterprise architects should look at data, technology and strategy as a whole to develop perspectives in line with all enterprise requirements.
That’s right. When it comes to technology and governance strategies, policies and standards, data should be at the center.
Strategic Building Blocks for Data-Driven EA
The typical notion is that enterprise architects and data (and metadata) architects are in opposite corners. Therefore, most frameworks fail to address the distance.
At Avydium, we believe there’s an important middle ground where different architecture disciplines coexist, including enterprise, solution, application, data, metadata and technical architectures. This is what we call the Mezzo.
So we created a set of methods, frameworks and reference architectures that address all these different disciplines, strata and domains. We treat them as a set of deeply connected components, objects, concepts and principles that guide a holistic approach to vision, strategy, solutioning and implementations for clients.
For us at Avydium, we see the layers of this large and complex architecture continuum as a set of building blocks that need to work together – each supporting the others.
For instance, you can’t develop a proper enterprise strategy without implementing a proper governance strategy, and you can’t have an application strategy without first building your data and metadata strategies. And they all need to support your infrastructure and technology strategies.
Where do these layers connect? With governance, which sets its fundamental components firmly on data, metadata and infrastructure. For any enterprise to make the leap from being a reactive organization to a true leader in its space, it must focus on data as the driver of that transformation.
DATA-DRIVEN BUSINESS TRANSFORMATION – USING DATA AS A STRATEGIC ASSET AND TRANSFORMATIONAL TOOL TO SUCCEED IN THE DIGITAL AGE
Data-Driven Enterprise Architecture and Cloud Migration
Let’s look at the example of cloud migration, which most enterprises see as a way to shorten development cycles, scale at demand, and reduce operational expenses. But as cloud migrations become more prevalent, we’re seeing more application modernization efforts fail, which should concern all of us in enterprise architecture.
The most common cause for these failures is disregarding data and metadata, omitting these catalogs from inventory efforts, part of application rationalization and portfolio consolidation that must occur prior to any application being migrated to the cloud.
Thus, key steps of application migration planning, such as data preparation, master data management and reference data management, end up being ignored with disastrous and costly ramifications. Applications fail to work together, data is integrated incorrectly causing massive duplication, and worse.
At Avydium, our data-driven enterprise architecture approach puts data and metadata at the center of cloud migration or any application modernization or digital transformation effort. That’s because we want to understand – and help clients understand – important nuances only visible at the data level, such as compliance and privacy/security risks (remember GDPR?). You want to be proactive in identifying potential issues with sensitive data so you can plan accordingly.
The one piece of advice we give most often to our clients contemplating a move to the cloud – or any application modernization effort for that matter – is take a long hard look at their applications and the associated data.
Start by understanding your business requirements and then determine your technology capabilities so you can balance the two. Then look at your data to ensure you understand what you have, where it is, how it is used and by whom. Only with answers to these questions can you plan and executive a successful move to the cloud.
2 replies on “Data-Driven Enterprise Architecture”
Well said. Many years ago, the mantra everywhere in the DoD was the building of an Enterprise Logical Data Model. While at Air Mobility Command, we had built a 5000+ entity LDM with more than 25,000 attributes. Interestingly buried amongst this wealth of information was the WHAT, HOW,WHERE, WHO, WHEN and WHY of the enterprise. In the process of articulating codes and allowed values for reference data we were also defining process states, events, etc. By looking at the subcategory structures we were defining enterprise taxonomies. As EAs, we were able to expose a whole big operational part of the enterprise by simply mining the enterprise data model, though the people building the LDM (in more than 40 subject areas) were focused on the minutae and metadata needed to constrain data schemas oblivious to the EA implications of the large, curated, and governed data model.
Prakash,
I agree 100% with what you said… UNLESS you are looking at the WHAT, HOW, WHERE, WHO, WHEN and… yes!… the WHY… you can’t really build an EA that is all-inclusive.