Unlock Cost Savings with Freight Clusters–Now in General Availability

Last year, we introduced Freight clusters, a new, cost-effective Confluent Cloud cluster type—purpose-built for high-throughput, latency-insensitive workloads—such as observability data, batch pipelines, and AI/ML data ingestion. Since then we've been working with early access customers to take their workloads to production, and in doing so, have helped them achieve 90% lower infrastructure costs, while maintaining the reliability we all know and expect from Confluent Cloud.

Today, we are thrilled to announce that Freight clusters are generally available for Confluent customers using AWS. You can now spin up Freight clusters in Confluent Cloud, and immediately take advantage of their unparalleled cost savings, enabling you to onboard more data-intensive workloads to Confluent Cloud in a budget-friendly manner.

Delivering more cost-effective options with Freight clusters

Our experience has shown time and time again that the true benefits of real-time data streaming come when teams no longer need to make hard decisions on which data is real-time, and can instead default to streaming everywhere. When companies make this transition, they truly unlock the network effects of a data-streaming platform. However, this hasn’t always been easy, given the costs involved. Freight is the solution to the problem, allowing companies to move even their most data-intensive workloads, such as observability and data pipeline use-cases, cost effectively to streaming.

Freight clusters give our users the ability to trade some cost savings for use cases that may not require sub-100 ms levels of latency, and are up to 90% cheaper than self-managing Kafka. They’re perfect for relaxed latency use cases such as logging, monitoring, etc. 

Freight clusters are built on the next-generation of Confluent’s cloud-native Kafka engine, Kora, to deliver lower cost by trading off low latency.

Cost-effective cloud replication—the driving force behind Freight’s cost savings

We’ve said this many times, but it's worth repeating: the most costly part of running Apache Kafka at scale is the inter-availability zone (inter-AZ) networking and bandwidth costs. It’s up to 88% of infrastructure costs when self-managing Kafka.

To achieve the cost-effectiveness we wanted for Freight, we evolved our Kora engine to support a  “direct write” mode, where data is written directly to object storage services like S3, bypassing local storage, and avoiding replication on the Kora brokers. These clusters replace expensive inter-AZ replication with inexpensive direct-to-object storage writes—trading latency (from sub-100ms in our other clusters to up to a second or two with Freight), for significantly reduced network costs. Furthermore, data is written directly to highly cost-effective object storage. Freight also leverages Fetch From Follower to intelligently retrieve data from followers instead of leaders, across availability zones. This means Freight can completely remove cross-zone traffic costs, when using clients configured with zone-aware produce and consume configurations.

Freight clusters also offer improved cost savings and connectivity through a new networking option called Private Networking Interface (PNI). PNI represents a large investment in building a technology to interconnect VPCs that is easy to manage, secure by default, and avoids the complexities of problems like VPC peering—all while remaining cost effective at massive scale.

There’s a lot more to say about the next-gen technology we’ve put behind Kora’s new architecture—and how it’s the driving force that will allow us to deliver on this promise of cost-effective Kafka for any use case. Stay tuned for a more detailed technical blog post in the near future.

Auto-scaling up to greater than 30 GBps of total throughput

Freight clusters utilize the same Elastic CKUs (eCKUs) as Basic, Standard, and Enterprise clusters, allowing Freight clusters to auto-scale to the shape of your workload and lower your costs, by not requiring you to pay for more capacity than you need. This means Freight clusters are truly serverless—they scale up and down instantaneously, and they’re always perfectly sized for your workload. You save money by only paying for the resources you use, when you actually need them.

We’ve made sure Freight can support the demanding, high-throughput workloads it's meant for. Freight auto-scales from its minimum 2 eCKU cluster size, all the way up to 152 eCKUs (representing a total of over 30 GBps of throughput), and all the way back down in a truly elastic manner.

Get Started Today

We invite you to explore the capabilities of Confluent’s Freight clusters for yourself. Powered by the significant innovations in Kora’s architecture, Freight clusters are available in General Availability on AWS—click here to start onboarding. Additional cloud support will be rolling out later this year, and you can sign up to get updates here.

If you haven’t done so already, sign up for a free trial of Confluent Cloud to explore the new features. New sign-ups receive $400 to spend within Confluent Cloud during their first 30 days. Use the code CCBLOG60 for an additional $60 of free usage.*

Apache®, Apache Kafka®, and Kafka® are registered trademarks of the Apache Software Foundation.