If you thought monoliths only lived on mainframes, you haven’t considered the impact of integration.
According to our annual research, a significant percentage (57%) of organizations have progressed to the second phase of digital transformation. This phase is marked by digital expansion as organizations stitch together applications to create a digital workflow—what we more generally call a digital experience. This causes an increase in integration between applications. This growth is typically focused on a few, key applications that are responsible for orchestrating workflows across business functions. While there is a tendency to focus on those applications that directly implement a customer experience, every business domain will see the rise of applications that ultimately become critical to their digital experiences. These applications will rise to be the modern equivalent of a monolith.
Monolithic architectures are considered those developed using an "all in one" approach. Logic and data access are centralized within one application. These applications are generally associated with mainframes and legacy technologies, but any language and environment can be used to build an application based on a monolithic architecture.
What follows from business reliance on such architectures is a set of dependencies that impact a wide variety of concerns, including security, cloud usage, and even acquisition of talent.
This is due to the increase in integration—that is, an increase in the applications that depend on the monolith for data access and business process execution. A good example of this is in banking, where monoliths serve to manage financial accounts and become critical transactional backends to a multitude of new services. Payment processing, for example, is a relatively new service that depends on existing account applications to function. Each additional payment service expands the dependency on that application and makes it more costly and risky to replace.
These extraordinary dependencies are a key driver of business decisions. We found that nearly one in three (27%) have repatriated applications from the public cloud. The top reason for this decision (47%) was a "high level of application inter-dependencies between on-premises and cloud applications." This is a contributing factor to the secondary driver of cloud repatriation: higher than expected cost (43%). Integration implies the transfer of data, and data transfers are one of the hidden but significant costs associated with hosting applications in the public cloud.
The characteristic of high dependencies through integration is not peculiar to the application architecture. The need for consistent business logic and data stores is achieved by limiting opportunities to introduce errors. This results in key applications becoming an authoritative source of logic processing and data access.
Interestingly, modernization increases the use of integration because it primarily relies on APIs to extend access to logic and data to modern components. To wit, a survey of IT decision-makers in 2019 found that "48% say modernizing existing IT systems and apps is one of their biggest integration challenges."
Access to consistent data is also critical to the success of an application, regardless of the architecture. O’Reilly noted in its research on microservices adoption, "not using a centrally managed database with microservices tends to be associated with failure." The need to access consistent data also increases the number of dependencies on key applications.
As a result, any application—regardless of architecture—might become a "modern monolith" by nature of this tendency to attract integrations (workflow gravity) for logic processing and access to data.
As organizations progress on their digital transformation journey, they expand their digital capabilities. These capabilities are expressed in the form of digital workflows, which are implemented via the integration of multiple applications and systems. This integration today is primarily accomplished via APIs.
This results in one or two key applications becoming the focal point of a workflow. An IDC study of digital transformation indicated that of the 81% of enterprises deploying automation, 65% rely on "workflow software that is embedded in another application, such as enterprise content management, content collaboration, capture, e-signature, etc."
Over time these applications become “modern monoliths.” Irrespective of their underlying architecture, they are a monolith in the characteristics that matter to the business: they are too costly and risky to replace due to extraordinary integration. They become as intractable as the traditional (mainframe) monoliths that continue to act as the foundation for a significant percentage of enterprise organizations.
Platforms—these modern monoliths—are increasingly strategic to both the business and the CIO tasked with executing on digital transformation.
We have hindsight, today, that enables us to recognize the rise of modern monoliths. Business and IT need to be aware of the long-term ramifications of standardizing on such a platform, and plan accordingly.