Data Modernization powered by APIs

Introduction

If your organization is embarking on a Data Modernization (or Application Modernization) process, and it is scheduled to last for some time, you have many challenges like: Keep current business processes, ensure an undisturbed transition phase, and have a stable Services platform post-migration. Your organization can leverage APIs to continue producing value; while helping the transition to new systems and data models. The Content Based Routing (CBR) architecture, in which a Router API can support different systems and data models by configuring rules, can mitigate these challenges.?

Current state takeaways

Your current state may present some issues to transition to the desired future state. These issues include Lack of data governance, low trust in data, and lack of visibility of integrations in place. I will focus this article only around “Lack of Data Governance”, where the following issues may arise:?

• Different data models (or metadata); absence of a unified data model.?
• Different systems as source of truth (master data); data needs integration.?
• Data aggregation from different systems/repositories to enable business operations.?

How Content Based Routing enables Data Modernization

• A Router API is deployed between the Main API and all Internal APIs.?
• Main API is diverted to call only Router API.
• Router API handles all requests from Main API with the following sequence:

1. Router (built with backward compatibility) supports Main API; mimics all existing entry points of the internal APIs; implements all operations and supported data models (<A>, <C>, <Public>; <UNIFIED> is added when is defined).?
2. The Router API inspects a request message and, once identified, uses routing configuration to determine its destination(s) and the transformation(s) needed.
3. Router sends the request to destination(s) selected with the proper data model.?
4. Internal API(s) responses may be aggregated (or filtered accordingly) and response is built with the corresponding data model.?

Benefits

• Support different data models (and operations) and one or more destinations/source of data.?
• Maintain legacy systems that its transition will take months to decommission (up to years).?
• Support of different sources of truth.?
• The route configuration can be changed to stop using destinations by changing the route configuration.

Industry Applications

• All industries can get benefits from this pattern, and its implementation can be sped up by using entry point definitions (like APIs specs) while those entry points are being built.?
• All ERP Implementations that replace a current ERP system.?
• Internet of Things (IOT) can use this pattern to support old devices while new devices with new bios and operations are being released into the consumer market.?

Conclusion

Your Data Modernization Program can invest in implementing Content Based Routing to address these Takeaways, your API platform will allow your organization to continue producing value, smooth your business transformation and adapt services to new business demands. The Content Based Routing Pattern is applicable to a wide range of industries where different data models are used.?

Takeaway: You can use Generative AI (GenAI) to build data models and their metadata!

Annex

Content Based Routing pattern (CBR)

What functionalities does this pattern need?

Entry Point?

• Mimic all desired entry points (operations and protocols) that internal APIs currently have.

Routing rules

• Set of configurations that indicate where to send requests, transformation to execute and aggregation when needed.
• A status configuration can enable or disable routes, and a date can also be used for the same purpose.
• More than one destination can be there, and aggregation can be executed to build a response.

Transformation

• In other words, it involves the transformation of data models from new data models to previous versions or vice versa.

Aggregation

• The function handles responses back to the Main API from only one system or from multiple systems or gives priority to one system over others

Challenges to keep this pattern working ?

• Any iteration over Router API must not break communication (contracts) with the Main API; Backwards compatibility.?
• Existing data models must always be supported.?
• The latest versions of Internal APIs must be welcomed.?
• Route API must be light and with high performance (scalable); enough instances must be available to allow traffic without affecting Service Level Agreements (SLAs).?

?