About the client:
The client is one of the largest integrated telecom providers in the United Kingdom, serving millions of customers across consumer, business, and government sectors. With a legacy of innovation and a nationwide infrastructure, the organization delivers a wide range of services including broadband, mobile, fixed-line, and IT solutions.
As a critical enabler of the country’s digital ecosystem, the client operates at massive scale and complexity, with stringent requirements around reliability, security, and compliance.
Benefits Overview:
40% Cost Savings
60% Faster Time-to-Market
Transitioned 5 Million+ Lines of Code
Challenge
The client operated over 90 mission-critical applications built on EJB 3, deployed on WebLogic Server. While these systems were foundational to day-to-day telecom operations, they had become increasingly costly and risky to maintain.
Key challenges included:
- High Total Cost of Ownership (TCO): Legacy infrastructure and licensing contributed to escalating operational expenses.
- Security Issues: Outdated technologies exposed the environment to growing security vulnerabilities.
- Migration Risk & Knowledge Drain: The complexity of EJB 3 systems, coupled with sparse documentation and attrition of legacy knowledge, made modernization efforts risky and error-prone.
- Big Monolith: Hard to manage a monolith with bad code design.
The client’s reliance on a dated but critical platform constrained innovation and introduced substantial risk. Traditional modernization approaches proved too slow and resource-intensive, raising concerns around service continuity and long-term scalability.
Solution Architecture:
Legacyleap led the modernization of the client’s EJB 3 system, leveraging Gen AI models to tackle the lack of documentation and the system’s complexity. The project began with automating code comprehension and business logic extraction, dramatically reducing the time and effort required to document and understand the legacy codebase.
AWS services complemented this effort by enhancing functionality and scalability. Amazon Bedrock’s Gen AI capabilities powered the analysis and documentation process, while Amazon Neptune was utilized as a vector database to map code relationships, ensuring smooth transitions during modernization.
By integrating Gen AI automation with a well-structured modernization framework, we delivered a high-performing, maintainable system architecture that addressed the client’s immediate challenges and positioned them for future scalability.
Legacyleap delivered a structured and AI-accelerated modernization solution, leveraging its proprietary code parser and Intermediate Representation (IR) framework for precise, scalable transformation.
The platform helped in breaking down the monolith to microservices architecture by analyzing the domain model and creating bounded contexts, which was later refined by domain experts, before performing the modernization.
1. Proprietary Parsing & Deep Analysis
- Legacyleap’s in-house EJB3 parser ingested .java files, dissecting complex API Endpoints, business rules, and integration mechanisms.
- Detailed analysis of API lineage, meta data, transformation logic, and operational dependencies was performed to capture the “as-is” state with complete traceability.
2. Assessment & Comprehension
- Legacyleap auto-generated detailed technical and business documentation, including flow diagrams, lineage reports, and component-level specifications.
- Technical Debt & Complexity Assessment: Highlighting modernization hotspots and optimization opportunities
- Transformation Readiness Report: Estimating migration effort and risk
This ensured knowledge preservation and simplified future maintenance and enhancements.
3. Microservices Recommendation
Legacyleap platform analyzed the existing API endpoints and the domain model, and based on the bounded contexts, it recommended microservices along the visual representation of architecture.
4. Automated Code Transformation
- Extracted Micro Services using recommendations. Each microservices was created as an individual project with all the characteristics of a 12 factor application.
- Leveraging the IR, Legacyleap’s code generation engine produced optimized Spring Code using Spring semantics like repository patterns etc.
5. Validation through Auto-Generated Unit Tests
To ensure functional accuracy, Legacyleap generated unit tests and API Endpoints for each class, covering:
- Business rule validations
- Edge cases & quality checks
- Regression scenarios
- BDD-style Gherkin test cases
Automated test cases ensured functional parity between EJB3 and Spring, safeguarding functionality and business accuracy.
Results:
40% Cost Savings
Reduced tooling, effort, and rework through guided automation.
50% Faster Time-to-Market
Accelerated dev with 70%+ automated code and logic extraction.
Seamless Transition of 5M+ Lines of Code
Re-architected EJB3 to Spring with full logic and API fidelity.
Enhanced Performance & Scalability
Modern Spring stack enabled faster APIs and cloud scaling.