Case Study: Samsung achieves a scalable IoT data platform and 60% infrastructure cost savings with Mesosphere (Mesos & Marathon)

Samsung is powering the Internet of Things with Mesos and Marathon

Samsung’s SAMI is a cloud platform that consolidates sensor and device data to help developers build intelligent IoT applications. Facing rapid growth in microservices and heavy data-processing needs, the team needed a way to scale containers, simplify device/cloud connectivity, and let developers and QA push code to production without managing underlying infrastructure.

The team adopted Mesosphere’s DCOS (Mesos, Marathon, Chronos) and Docker, running Jenkins and Spark as Mesos frameworks on a shared public-cloud cluster (about 500 CPU cores/800 GB RAM in production, ~400 containers). This architecture enabled >50 microservices and large batch jobs with easier CI/CD, eliminated separate Spark masters, improved resource predictability, and cut infrastructure costs by roughly 60%, allowing frequent daily releases and more efficient scaling.

Samsung

Jerome Dubreuil

Senior Director of Engineering