How To Automate an API Gateway Migration
In this article, readers will learn the best way to simplify an API gateway integration and migration to abstract API complexity and apply API automation.
Join the DZone community and get the full member experience.Join For Free
Application Programming Interfaces (APIs) and API management are critical business assets. Today’s businesses are scaling APIs to meet different business needs with varying types of APIs, specifications, and functions across enterprise environments and complex architectures. Smart developers understand the business value of APIs and the need to have an API strategy independent of API gateways and host locations.
Any workable API strategy must include gateway-agnostic APIs so they can run on-premises, in the cloud, or in a hybrid environment. APIs also must be easy to port to any new gateway provider. Whether the organization uses APIs custom-built by the DevOps team or purchased as Software-as-a-Service (SaaS), using APIs to connect applications throughout the organization requires effective API management. As the speed of business accelerates, APIs need to move faster as well, which means APIs need to be unified and self-service to be more agile and more portable.
API lifecycle management must be able to accommodate API migration. API management tools enable organizations to get the most out of their microservice platforms, including managing API gateways. While there are standardized API structures like SOAP (Simple Object Access Protocol) and REST, there isn’t a standard set of best practices for API management. It also means that migrating from one API gateway provider to another can be highly complex and costly.
The best way to simplify API gateway integration and migration is to abstract API complexity and apply API automation. Using a low-code/no-code approach, you can build and manage APIs through a single, unified platform that is readily portable if you must change API gateway providers.
The Challenge of API Gateway Migration
From a technical standpoint, an API gateway isn’t essential. You can hire a programmer to create an embedded code response for call request routing. Of course, that will require additional staffing and resources to develop code for individual call requests. An API gateway eliminates the need for extensive recoding and the risks and errors of programming individual connectors. It also makes migration easier since the built-in connectors are already in place.
If you have a single API management tool that provides end-to-end lifecycle management, it eliminates the need to use multiple API platforms. It also unifies the process for all stakeholders, including support for testing, compliance, security, etc. You have an API management solution that functions independently of specific gateway providers.
Consider the challenges of migrating between API platforms. For example, say your company wants to migrate from Apigee API Management for Google Cloud to Kong API Management. Without an API gateway, a lot of time and resources would be needed at the outset to assess gateway configurations. You would start by reviewing the current API gateway configuration. Next, you would have to assess each gateway to understand how migration might work, what needs to be extracted, and where it needs to be routed. Implementing the migration using manual programming would require time, resources, and extensive testing to validate that the migration was successful.
Preconfigured Connectors Simplify Migration
The migration process is much simpler when you have an API management platform with built-in platform-agnostic connectors. Consider how you would approach our Apigee-to-Kong migration example using a low-code platform that decouples the control plane from the gateway:
You begin by configuring the connectors, adding Apigee connectors based on two variants:
- For OPDK and Edge, you configure using Org name, User Account details, and deployment type (on-premise or legacy Edge); or
- You use Apigee X for newer Google Cloud-based SaaS offerings using a configuration with Org name and JSON file with JSON Web Token (JWT). You then add the Kong connector using a configuration that includes the Org name and JSON file with JWT.
- Next, you analyze the deployment, starting with your current Apigee deployment and how the APIs are consumed. You need to map the details of the apps, policy, target, and associated resources and determine what is part of the API proxy and its dependencies.
- Now, you must map the policy/functionality/plugin between Apigee and Kong.
- The final step is to review all the configurations and edit the mapping of services and routes as needed. You are now ready to initiate migration.
Taming API Sprawl
Adopting a low-code, unified approach to API management simplifies API gateway migration and eliminates other developer headaches.
Using API automation tools allows you to preconfigure connectors, simplify and standardize coding, generate documentation, build a catalog, and push it into the delivery portal. Automaton eliminates much of the complexity that plagues API development while making it easier to adapt and migrate APIs as needed.
Experts predict that, by the end of 2022, 90% of new digital services will consist of composite applications that use both internal and public APIs on-premises and in the cloud. With the growing API sprawl, ensuring the API lifecycle does not depend on the platform, or provider, your choice is more important than ever. Using a decoupled API control plane simplifies the development of new APIs and API gateways and streamlines API gateway migration.
Opinions expressed by DZone contributors are their own.