Kafka has been a crucial facet of the overall SAS Customer Intelligence 360 (CI360) architecture for quite some time. Until 2021, Kafka supporting CI360 was managed on standalone virtual machines. Traditional VM backed infrastructure posed administrative challenges for ensuring consistent software patching, adding scale on demand, and providing a highly available, redundant, and durable message bus for the CI360 microservices.
The goal was clear, the backend Kafka platform needed to move from the aging legacy systems to a more cost effective and stable solution.
The standalone VM backed Kafka clusters were migrated to the Amazon Elastic Kubernetes Service (EKS) with zero down time. Cluster Linking and the Confluent Operator were used as part of this effort. Both technologies were crucial in ensuring that the systems were online and available throughout the migration.
This session details the journey for moving standalone Kafka to Kafka on K8S. During the session, scope of the journey including Total Cost of Ownership (TCO), technical architecture, and the migration itself will be discussed.
NOTE: Experiences related to this effort are being published in a joint case study between SAS and Confluent titled, ""SAS Powers Instant, Real-Time Omnichannel Marketing at Massive Scale with Confluent's Hybrid Capabilities"".