Confluent Platform enables unparalleled Kafka performance and elasticity through automated partition rebalancing, decoupling of the compute and storage layers, and infinite data retention.
Self-Balancing Clusters automate partition rebalances to optimize Kafka’s throughput, accelerate broker scaling, and reduce the operational burden of managing a large cluster. Partition rebalances are completed quickly and without any risk of human error.
Tiered Storage allows Kafka to recognize two tiers of storage: local disks and cost-efficient object stores (Amazon S3 or Google Cloud Storage). Brokers can offload older topic data to object storage, enabling virtually infinite retention.
Self-Balancing Clusters monitor the skew of resource utilization across your brokers and continuously reassign partitions to optimize your cluster’s performance and balance.
With Self-Balancing Clusters, you can add or decommission brokers and a partition reassignment will automatically be triggered, providing you with even greater cluster elasticity.
Eliminate the need for complex math and the risk of human error that manual partition reassignments typically entail.
Achieve dramatically longer periods of data retention for your cluster without a significant increase in operating costs. By leveraging Tiered Storage, you can back up data for replay in the future or use it to make Kafka the central nervous system of your organization.
Tiered Storage also allows you to achieve the regulatory compliance for data retention and durability requirements specific to your industry, without needing to build additional infrastructure into your architecture.
Scale storage resources without having to scale compute resources, and vice versa, by effectively decoupling compute and storage resources allocated to your cluster through Tiered Storage.
Minimize the time needed to rebalance partitions when adding new brokers. Because less data is stored on the broker when using Tiered Storage, smaller partitions need to be moved to newly added brokers using Self-Balancing Clusters, reducing rebalancing times from potentially hours down to a few seconds.