Client
One of the largest hotel chains in the world, the client has 30 brands and 7,000+ properties across 131 countries and territories. Taking hospitality beyond what was thought possible, the client has over 170,000 employees across the world and generates billions of dollars of revenue each year.
Challenge
One of the main challenges for this project was the adoption of more modern technologies, considering that the client was running their core business-critical process on legacy mainframe systems. The mainframe implementation was one of the core sources of compute and data for a lot of downstream business critical applications including transactional front-end and critical analytics applications. But the advantages that resulted from it also resulted in several issues that affected key business metrics, such as the high cost of running the mainframe, both due to its scale, as well as the storage required. Not only was maintaining it an issue, the talend that would be required to do so was getting increasingly harder to find, considering that the technology was old. This also meant that new technologies were not possible to integrate, thus affecting usability and performance both. Moreover, the existing integrations were implemented using monolithic APIs, which needed to be fixed as well.
Solution
The process for cloud adoption started with detailed consulting and planning discussions focused around design and architecture to understand the current state of mainframes for the client. Moreover, there was a discussion on how to transition to cloud with minimal downtime. At the next stage, Mindtree created technical PoCs to validate the workload movement to cloud and to preserve the integrity of the integrated systems. For this, AWS was used as the cloud platform, Kafka and Spark were used for data transformation, Couchbase and Postgresql for data storage, and Hadoop as the data lake platform.
As a result of this, we enabled better real time and batch ingestion of data from multiple sources into the data lake and were able to transform monolithic implementation of APIs to microservices-based implementation without breaking the front-end analytics integrations.
Benefits
By transforming the mainframe to cloud, Mindtree was able to deliver the following benefits:
- Reduced operational costs by eliminating the mainframe workload, storage and maintenance costs
- Reduced operational overheads and simpler to maintain
- Cloud agnostic architecture allowing for future cloud platform movements
- Improved flexibility on the front end through microservices-based architecture
- Opportunities for optimizations and improvements in the analytics and insights space of the application
