SDTimes has declared 2020 the year of integration and is featuring an article with Red Hat’s Sameer Parulkar focusing on API-first design, agile integration, and distributed architectures. This approach is something that Red Hat has been championing since 2017—as technologies and applications have evolved, integration architecture has had to evolve to keep pace.
But it’s worth taking a step back and asking—why is this the year of integration? Why now?
The SDTimes article provides some hints on why integration (a first-tech bubble technology concept) is so critical now. They list a host of digital transformation-related initiatives, such as IoT, streaming data, data warehousing, containers, and cloud. All of those both produce and consume massive quantities of data and rely on strong interconnectivity to be effective.
Conceptually, things like IoT and microservices aren’t new—but they have hit levels of maturity and adoption in the last few years that were simply not possible a decade ago. These technologies are the core of digital transformation initiatives. As both business and IT managers have focused more on digital initiatives, they are encountering an ever-growing need to be able to view, manage, and connect data associated with the new transformation projects. That’s the high-level view of integration.
Digging a bit more deeply into what is meant by "integration," though, and it becomes clear that integration itself is evolving and maturing in tandem with other technologies. The SDTimes article calls out that modern approaches to integration blur the lines between integration, application development, and data analytics (which is probably the result of trying to develop microservices, data pipelines, and SaaS or cloud applications, and then connecting them all).
And that circles back to Red Hat’s embrace of agile integration.
As a quick definition, agile integration is an architectural and process approach to integration that parallels agile development or DevOps by taking a cross-functional, end-to-end perspective on what is being integrated and the technologies to do it.
What that means in practice is that an agile integration approach relies on some technologies to enable rapid development and CI/CD pipelines, namely using containers for developing and deploying any kind of integration definition (from a Camel expression to set of APIs). From there, Red Hat proposes having a "toolbox" approach to integration platforms -- rather than having a single integration technology to try to address every application or data need, having a set of very different types of technologies available provides the ability to use the best type of integration for a given scenario.
This agile integration architecture is well suited to distributed architectures and technologies like IoT.
So, what does this mean for the upcoming year of integration?
Some of the biggest reasons that customers are moving toward agile integration as an approach is because of how embedded hybrid cloud and container architectures are becoming, according to Sameer Parulkar. These distributed environments (and applications) require distributed integration.
As part of that, Parulkar pushes for an API-first approach to application and integration design. APIs are simple and more secure ways to cross cloud boundaries while maintaining access to data and to applications. API-first design puts the focus on the ways that users or applications will connect and what their needs will be, which helps to keep the integration itself focused on the right kind of data to access and the methods to access it. Using an API management platform adds in additional layers of security around API publishing, user authentication, and traffic routing.
Parulkar also pointed out another advantage to defining integrations within Red Hat OpenShift: using Operators. Openshift 4 introduced the idea of defined services which can be configured and deployed in specific situations through OpenShift. Different types of integrations can be set up as these Operators, and then configured and deployed where needed, rather than having to create a new application in another container and deploying it.
Effective integration strategy requires a lot more than a technology platform. It requires a design vision, process and cultural change, and (still) a toolbox of technologies that can address different scenarios. This kind of agility allows IT departments to change directions and priorities and to introduce new initiatives on a consistent foundation.
And as SDTimes notes, integration is crossing more boundaries than before, touching application development and design and data science -- crossing the worlds of both I.T. and business leadership. Having an agile integration approach can enable I.T. leaders to collaborate and deliver on business priorities, which is the real-world effect of digital transformation.