At Amadeus, the BRMS is a key enabler for the customization and the configuration of our products2019 functionalities to answer the diversity of our customers2019 business requirements: Airline companies, Travel Agencies, Hotels, Airports, Rail companies. Amadeus C++ and Java application servers handle several million transactions per second. Knowing that each transaction performs several business rules evaluations, it is crucial that the rule engine framework is as efficient as possible with respect to the latency, CPU consumption, memory footprint, and scalability. As a middleware team, our responsibility is to offer the best possible rule engine solution that answers both our functional and non-functional requirements. We selected Red Hat BRMS/Drools and have worked on filling the gaps to make it fit Amadeus architecture, performance and operability needs; the key challenges being: integration in Amadeus C++ stack, under-millisecond response time, 0-outage rule updates. In this talk we2019ll present how, by leveraging on the collaboration between RedHat and Amadeus, we2019ve been able to fulfil these requirements and how we envisage to upstream this development to the BRMS community.