Dan Temkin is a Senior Product Manager at IBM Automation. In this article, Dan discusses unleashing business agility with API-led integration.
We have seen some challenges in the marketplace over the last few years.
- The great digital shift – In the last year, we did ten years of digitization projects.
- Growing complexity – 80% of organizations have over 1000 applications in their portfolio today. We want to ensure that all those different applications can be accessed as APIs.
- Attrition skill gap – Worldwide, we had 4.3 million people quit in August; it increased to 4.5 in November 2021. Organizations are saying that they are trying to do a lot more, deliver more assets and functionality, and have greater agility, but with fewer people. Often those people are more general purpose; they are not always a core skill set group.
IBM Automation Strategy
- People – Combine human skills with automation and AI to enhance productivity and improve decision-making.
- Business – Digitize, integrate and orchestrate workflows and applications to drive growth, performance, and agility.
- Systems – Continuously observe, correlate, and remediate to assure cost-effective performance of applications and systems.
From an IBM perspective, we see that you don’t have automation without integration. Because at the end of the day, I want to be able to compose all these different assets together. These different assets could be part of your business that are exposed via services or could be services provided by other vendors. These assets could be legacy applications, new applications you’re putting in your environments, or SAAS-provided systems. Each one of these will have different interfaces; the way they describe the data and how they interact with their data will also be different. Some may use REST, SOAP, Async, or GraphQL. I should be able to do that all through a common API management platform. From our perspective, all of those different types of APIs fall into that same category.
Three characteristics slowing classic integration today
- The integration library and code base are bloated. So, when a part of that code needs to be changed, I will have to go through the whole testing process, the whole lifecycle process, and validation.
- Specialist-only dependency hinders scaling of delivery. So, we need more people with multiple and general-purpose development skills.
- Lack of interoperability and limited reuse.
API development is the solution to this. Two main things to focus on are –
- It should simplify software development
- It should be easy to do and secure.
A Simple Model – Creating a new API that registers business leads.
The business goal is to create a new API that registers business leads.
I have a few fields that I want to collect, first name, last name, etc., as part of the user profile. As I am following best practices, I need an ID as a primary key. I want to retrieve and update data. And I need to connect this API contract to our back-end enterprise lead management system. But I don’t want to necessarily have it where I’m going to just use their model and expose it right up to the front. Because I want to be able to change my lead management system, and I do not want to go back and rewrite all the calling applications that are going to be integrated with it.
After this, we test connectivity with each of the back-end systems.
Once the connectivity is working, I need to decide what data to expose to avoid leaks and make it useful for the users to consume the service.
Next is mapping data formats between the API and the lead management systems. Data mapping is tricky and takes a good amount of skill.
I also need to create test data because, eventually, I have to validate that the API works.
The API needs to meet security standards. It needs to be discoverable by the consumers and needs to be scalable. In addition, the customer may have additional requirements that need to be built in.
We want to deliver the functionality for the business efficiently and effectively.
IBM has spent a lot of time taking our different core capabilities and streamlining them into an integrated authoring experience. We have delivered the ability to build APIs in minutes driven from different back-end systems where I can start with a connection to Salesforce, SAP, or DB2. It allows consumers of the API to map fields enjoyably, allows the API provider to do this at speed and scale, and provide all the different functionality without having to write any code. It is a matter of a few clicks and checkboxes. It is developed using Open API Specifications and implemented using the latest security standards.
Some of the levers that we put in to provide accelerators are –
- Create integration flows using natural language.
- AI helps to select from known templates and existing connectors.
- Create first-class integration APIs in a no-code editor
- Catalog of hundreds of Smart Application Connectors
- Automatically generate OpenAPI spec and documentation.
- Natively supports model-driven REST API development pattern with a graphical model/schema builder and an integrated debugger.
- Graphical flow editor allows quick implementation of API or integration in minutes.
- AI transformation generator to build complex transformation logic from samples
Automation at every step
The process is a continuous feedback loop for continuous improvement
- Enterprise Architect – Designs connections between applications and data with a reusable integration process
- Integration Developer – Selects the best reusable integration pattern matching business needs.
- Integration Operations – Wants to automate issue identification and fixes for running integrations.
To conclude, to get more details, you can study API management use cases and industry examples at https://www.ibm.com/cloud/api-connect
You can also try it for yourself at