Click to learn more about author Romi Stein.
We now know the story of how, over the past decade, the CIO role has transformed to be that of a coach. IT has morphed into a democratized DevOps team, requiring a new model for leadership. A bevy of new coding practices, from DevSecOps to DataOps, has led to the increasingly sophisticated microservices that we use today.
Yet despite this cataclysmic evolution over a very short timeframe, the APIs that connect everything are still, unjustifiably, integrated by hand. It almost seems like integration was an afterthought, buried within the old IT pyramid. The costs to develop and maintain an API from digital sources can be between $7,500 and $12,000, with monthly costs as well. Considering that the average enterprise uses 1,295 cloud services, the costs add up quickly.
Integration is now the final frontier of IT democratization. The final member of the “-ops” family, IntegrationOps involves going beyond choreography, orchestration, or composition to automate API integration – and not a moment too soon.
A Cacophonous Orchestra
To get APIs to work together, organizations use choreography, orchestration, or composition. Choreography doesn’t involve middleware; it alerts all subscribers to a single data event and enables them to do what they want with the data. Orchestration uses middleware to wait for an API call, centralize data, and direct it to other APIs. Composition directs data to other APIs without requiring an incoming API call.
Contrary to their harmonic names, choreography, orchestration, and composition create cacophony between core systems and the cloud. Legacy systems generally have big code bases and incomplete documentation. They are still around because they’re resilient and cost-effective, but they’re also monolithic and weren’t built for interoperability in a world of cloud-native digital sources. Companies have thus accreted many layers of middleware in order to “bolt on” a digital facade and provide access. These include multiple points of API integration, web service wrappers, enterprise service buses (ESBs), business rule management systems, and other forms of customized tooling to provide compatibility with the cloud.
While cloud-native IT infrastructures can use an iPaaS (integration platform as-a-service) for drag-and-drop API integration, those with core systems have no such luxury. Instead, it can take months to integrate each layer of middleware. That is why integration is the final frontier of IT democratization.
IntegrationOps: The Final Frontier
In IntegrationOps, a command-line interface (CLI) provides rules of engagement and automates API integration. This enables anyone to spin up and orchestrate APIs without centralized IT. CLI – combined with modularity, low-code/no-code, and automation – make up the key components of IntegrationOps.
Why did it take so long for IntegrationOps to come to fruition?
It has to do with the nature of core systems and the difficulties in integrating with monolithic core systems. Additionally, big, traditional enterprises such as banks are in a 10-year migration process off of their core systems onto the cloud. From insurtech to DeFi, lines of business are converging more quickly than that.
In order to turn core systems into the composable units required for new lines of business, enterprises have two choices.
One is to buy one vertical solution for each type of core system: one for your IBM mainframe, another for SAP, a third for Oracle, and so on. If that’s not feasible due to the type of core system, or cost-prohibitive, the other solution is to find coders to manually wire APIs into their mainframes – a challenge of its own, given the scarcity of such talent.
IntegrationOps uses the modern concepts of software development – decentralized, no-code, automated code generation – to create an API factory. This factory sits as a digitally native layer on top of core systems and generates APIs based on digital representations of legacy assets. IT can use IntegrationOps to create specific integration components that deal with only one function, spinning up new integration in weeks instead of months, and dramatically cutting costs.
Implementing IntegrationOps
As demonstrated by the abrupt transition to remote work and the increased demand for digital services, the urgent need for IntegrationOps has only further been brought to light. However, this initiative cannot be implemented overnight and requires buy-in from multiple areas of the organization.
For management, it requires a change in culture and challenges an organization to establish channels of authority and ownership around technology. By establishing a steering committee composed of both business and IT leadership, organizations can ensure that their implementation of IntegrationOps is effective.
When it comes to the technology side, teams can begin their approach by mapping out their core systems, as well as the associated technologies built to access and interoperate with them. Guided by the principles of IntegrationOps – automation, standardization, and modularity – they can then evaluate their current “stack,” prioritizing the existing bottlenecks and upgrading tools where the seamless development of digital services from core systems are most hindered.
Fully Achieving Digital Transformation
IntegrationOps essentially removes the final barrier to digital transformation. For many enterprises with core systems, the technology arrives just in time. Banks, health care organizations, and anyone housing data now face new regulations and interoperability requirements. Because IntegrationOps sits as an entire layer over core systems, it can capture screen, data, and logic in one place, offering flexible access to compliance demands. The biggest problems of the integration stack are essentially solved, and IT can consider itself democratized.