New Challenges for Integration
Integration is not Legacy! It's a living world with a thriving ecosystem of developers who provide answers and solutions to actual challenges.
Join the DZone community and get the full member experience.Join For Free
I often hear that integration is an old subject, which does not move anymore, with old, has-been technologies; in short, that it is a non-subject. Those who think that have not necessarily understood that a topic like API Management is a subject of integration and that integration has reinvented itself. I wanted to share with you what I felt were the challenges facing the topic of integration right now. Do not hesitate to share your comments and opinions on the current challenges faced by integration.
Cloud and IoT Integration
Let me start with this: IoT integration is nothing more than cloud integration. By this, I mean that most developers who integrate with IoT, do nothing but integrate an IoT platform, and see little intrinsic complexity of the IoT devices. So, if we include IoT, SaaS solutions, PaaS solutions, and so on, we are faced with problems common to all these platforms. For example, their very high elasticity, or their ability to transfer very large volumes of data in a limited time. These cloud system capabilities can hardly be found on-premise and therefore also require cloud integration platforms. We potentially end up with two dedicated platforms. One "mature" platform in on-premise, and another, cloud-specific, platform. The whole issue then lies in the right definition of trajectory making the two platforms coexist, to see them merge.
I know, we quickly put Agile everywhere, even placing modular and decoupled words without really knowing what that means. To fight against this, I can only recommend my two reference books, "The Agile Architecture Revolution" and "Building Evolutionary Architectures," which can really help you to understand the underlying problems of an IT system capable of evolving rapidly. To that end, I would add a certain vision of architectures, i.e. grouping and classifying architectures, according to four big families:
Integration architecture by data.
Integration architecture by services.
Integration architecture by events.
Integration architecture by process.
All these types of architectures continue to live on, as we have seen with the increasing popularity of event-driven architectures, including the appearance of event-oriented APIs. In fact, I do think that Integration will continue to live as long as new technologies come out.
Integration as a Product
Integration is increasingly becoming a product:
Setting up co-construction mechanisms inspired by what is done in microservices projects, which are themselves inspired by open-source projects.
Setting up a complete and automated production line.
Offering monetized service catalogs via API Management.
This product logic comes directly from the agile world, so you will know what to say when someone tells you that integration is a has-been.
Two phenomena can be seen between microservices and integration. The first is that without microservice point integration if we want two microservices to communicate with each other, it is already a matter of integration. As a result, new solutions are emerging that will only enhance the importance of integration. The other phenomenon is that microservices alone in their corner are not always legion, and they must be integrated into the Legacy world. To do this, microservices and integration have to interact with each other, and interface as well as possible.
What about you? How do you feel about that? What are the new integration challenges you see? Feel free to exchange with me in the comments.
Opinions expressed by DZone contributors are their own.