NatCo of a Leading European Telecom
Telecom mobile applications and other data consumers constantly retrieve data from CRM and other transactional databases risking overload. The goal of our client, a NatCo of a leading European telecom, was to reduce the number of requests to operational systems and latency to below 300ms.
We built a data model to integrate customer and product data from source systems in real time. Loading in real time varied among source systems depending on the underlying technology. Kafka was used for streaming, and KSQL for data transformation in the streams. A caching system was used for all open sessions propagating changes back to the source.
Mobile app latency was reduced far beyond expectation. The number of requests to operational systems was minimized. The architecture was implemented as a dockerized stateless service, and the REST service for communication with data consumers was built on top of it.