Bridging Old and New: Designing APIs That Connect Legacy Systems with Modern Microservices

Nowadays, choosing between legacy systems that have served the company for years and modern systems is a complex decision. However, instead of making this choice, integrating them would be more financially and operationally beneficial. It would reduce costs and accelerate time‑to‑market without discarding crucial business logic. We invited Sumit Saha, a Software Engineer from a BigTech to shed more light on this.

1. The difference between legacy systems and microservices

Sumit mentions, Legacy systems, such as on-premise ERP systems that do not support modern cloud ERP or legacy LMS platforms at universities, often fail to meet the requirements of current operational processes since they are not compatible with new mobile features. These requirements include agility, scalability, cloud integration, and REST API integration. Although legacy systems continued to perform reliably for years, the modern market requires faster deployment capabilities.

On the other hand, microservices are built around small, independently deployable services. Each addresses a single feature, communicated over new coordination protocols (for example, REST, JSON, or gRPC), and is integrated with CI/CD, observability, and container orchestration. The difference between them and legacy systems is that the former supports faster feature delivery and scalability. However, it often happens that microservices alone lack the domain logic that legacy systems had. Therefore, bringing old and new ecosystems together to achieve the best business results becomes essential.

2. Why integration matters

He also suggests that, integration matters for two main reasons:

1. Preserving valuable systems performing

Besides business logic, legacy systems often embody domain rules, compliance knowledge, and operational history. Replacing this information is costly and risky. Fortunately, modern APIs allow organisations to modernise what is needed while keeping legacy systems unchanged.

2. Integrating new services

Instead of rebuilding systems from scratch, organisations can release new services by layering microservices on legacy systems. For instance, a bank might launch a new online payment API by adding an API layer on top of the core system, which significantly reduces time to market.

3. Integration challenges

It is clear that bridging old and new systems together is highly beneficial; yet according to him there are challenges to consider:

Data silos

Old data is stored in proprietary formats. Converting it to, for example, JSON, requires additional cleaning and validation.

Protocol and semantic misalignment

COBOL's record structured and two-phase commit transactions do not align with RESTful JSON or asynchronous messaging. In this case, translating the form and semantics, such as compensating logs or transaction boundaries in the case of a bank, requires middleware or API generation tooling.

Security risks

Legacy systems were not designed with a modern API security perspective, and exposing them can lead to compliance breaches. Hence, it is necessary to implement authentication, authorisation, and encryption to protect sensitive data (for example, transaction history).

Performance constraints

Legacy platforms may not be capable of handling the high velocity of API calls generated by modern frontends. Proper API management (rate limiting, throttling, and caching) is needed to solve the issue.

4. Case studies

There are a few examples of successful API-based integration strategies. As such, in the finance services sector, Bank Leumi, one of Israel's largest banks, modernised its operational processes to become an API-first organisation. The company used automated tools to generate APIs from COBOL mainframe applications and exposed its legacy functions as microservices. This approach enabled compliance with open banking regulations, reduced infrastructure costs, and promoted digital innovation without replacing the old system.

In the healthcare sector, hospitals are modernising legacy electronic medical records (EMRs) by integrating them with cloud-native platforms, such as Epic-as-a-Service. However, rather than migrating the entire system at once, many hospitals first used legacy APIs to expose patient data in Fast Healthcare Interoperability Resources (FHIR) formats. As a result, organisations modularly upgraded their systems, including digital appointment scheduling, billing, and telehealth services.

5. Tips for successful integration

Sumit suggests that it might be helpful if your organisation plans to integrate legacy systems with modern microservices with these tips:

Create consistent APIs

Use consistent naming conventions and standardised data formats (JSON). Depending on the use case, adhere to the RESTful concept or gRPS.

Focus on a single feature

Each API should expose a single function. Avoid overloading the endpoint with too many responsibilities.

Use API management tools

Apigee, Kong, or AWS API Gateways help with throttling, monitoring, and analytics.

Use observability

Monitor API traffic, latency, and errors using logging and dashboards to identify and resolve issues quickly.

Conclusion and the future

Bridging legacy systems with modern microservices is a strategic business advantage, allowing organisations to protect their core systems and implement the agility and scalability of modern software simultaneously. As for the future, this trend is expected to continue. According to a survey, most of the participating companies are satisfied with the results of their cloud migration, and almost all report increasing their budgets for cloud technologies and integration. These statistics show that bringing old and new systems together becomes relevant, and companies should consider implementing this practice.