Don’t Let Bimodal IT Inhibit a Real-Time Architecture

As more and more companies go through their digital transformation journeys, many are adopting some form, either formally or informally, of what Gartner calls Bimodal IT. In a nutshell, this approach advocates executing IT delivery using separate approaches—one focused on stability and predictability (Mode 1), the other on speed and agility (Mode 2).

While there are many differing opinions on the long-term effectiveness of this model, I’d like to focus on some of the impact it has on enterprise architecture. Specifically, how the technical approaches used to integrate across the modes can hamper the ability to have a real-time architecture.

As I work with different companies that have separated their digital teams from the rest of IT, I have begun seeing a common architecture emerge in many integrations:

In the short term, this architecture makes sense. We have systems of record that are not capable, technically, of providing data to the systems of engagement for a variety of reasons. So an intermediary “data hub” is established to facilitate access to the data of record. This “hub” could be an RDBMS, data warehouse, big-data platform, in-memory data grid, or a combination of the several technologies. From there, the digital team can write specific services for their systems to consume, and possibly even alter, the data.

The problem with this architecture is not that it is bad. On the contrary, it is often the most viable solution for the initial stages of digital transformation. The problem with this architecture is that it is typically not envisioned, sold, and implemented for what it is: a transitional state, not an end-state. This design should only exist for the time needed to retrofit the systems of record so that they are able to fully participate in real-time interactions with the systems of engagement. Otherwise, the complexity and data latency inherent in this kind of architecture will continue to hamper the enterprise’s overall digital transformation.

Additionally, once this architecture has been in place for a significant duration of time, the main barriers to moving away from it can be as much cultural as they are technical. It can be difficult to get the two teams to trust each other in the implementation of real-time interactions. This is especially true of the digital (or Mode 2) team who like having that complete separation from the systems of record, so they don’t have to worry about the scalability and stability of those systems. The “hub” architecture offers them a kind of security blanket, providing the comfort of control over the producer of their required services.

But if an enterprise is ever to truly transform and completely cross the digital transformation chasm, the entire architecture will need to work cohesively in a real-time manner, free of the barriers that data replication and other legacy band-aids introduce. Whether or not the enterprise continues with a bimodal approach to delivery, the delivered solutions must work together seamlessly, reacting to changing business needs with agility and stability.