The client is one of the largest credit card issuers in the US market.
Many of client’s software systems are legacy mainframe applications. The client has started adopting microservices architecture and Agile development methodology and has established a DevOps-based culture.
These initiatives are undertaken in order to enable faster time to market and reduce the turnaround time for each release so as to meet the delivery timelines. Infosys has partnered with the client to cascade DevOps culture and re-engineer existing applications into event-driven, highly available microservices-based architecture.
Key Challenges
- Enormous monolithic applications built over 25 years across geographies and currencies
- Lack of documentation, regression test suite, and SME knowledge making it very tough to decipher the existing code base
- Non-Agile software development methodology and mindset involving long cycles of planning and release time
- Most of the existing systems are batch-based, making it tough for transition to event-based microservices
- Tight coupling within monolithic applications resulting in challenges in decomposing functionalities and making it hard to change any logic/functionality in it
Ready to experience?
TALK TO EXPERTS
The Solution
Leveraging Agile and DevOps Culture
- Infosys collaborated with the client development teams in order to establish architecture patterns for event-driven microservices
- Extensive training and self-learning enabled Infosys team to scale up quickly and become experts in various solutions like Spring Boot, Dropwizard, etc.
- Infosys partnered with the client and cascaded Agile and DevOps culture across the board
- The Infosys domain SMEs are learning new technologies and adopting Agile and DevOps-based mindset. This helped the client to not only explore the existing system but to also re-engineer the monolithic legacy applications into a modern microservices-based architecture by leveraging appropriate technology for suitable business cases
- Infosys helped establish components to convert record in files to event messages and vice versa, so that event-driven microservices could coexist with batch-based systems
- Provided recommendation to encourage documentation of learned domain knowledge and system designs in order to scale up the productivity of the Agile teams
Framework and Methodology Adopted
- Re-engineering to microservices architecture is much better than simple migration
- Testing (BDD and TDD) is very important. We conducted unit and integration testing for all the services written, to not only improve the development process but also help document the behavior of the components and avoid any future regression bugs
- Standardization of technology frameworks is overrated. It is better to choose the tool/frameworks suited for the given job
- Introduction of Kafka for asynchronous communication between end points
- Couchbase DB for handling large volumes. PostgreSQL for ACID compliance
- Used CI/CD pipeline for faster release by automating execution of unit and system and integration of test cases
Benefits
Transforming processes to simplify operations and enable efficiency
Microservices handling 5M requests a day with response time of less than 3 milliseconds per request
More than 250 Microservices were developed where in a few cases, services were re-engineered from main-frame components
Established CD / CI pipeline coupled with high Agile maturity, which has helped enable reduction of time to market and handle multiple deployments every day
Helped the client reduce the overall complexity and reduced cost of development, maintenance and opened the architecture for future extensibility
Behavior Driven Development reduced overall cost of quality and enabled faster release cycle
Change in mindset from "long cycles of planning" to "fail-smart & learn-fast"
Increased documentation enabled quality, interactive documentation and save time to publish and market the API’s
Enablement of on-boarding new comers
