Apache Kafka rebuilt for S3
We rewrote Apache Kafka from scratch on top of S3, resulting in some unique capabilities.
Never worry about Partitions again
Properly partitioning your topics is the most major issue running Kafka at scale. Your produce/consume workload may change at any time causing hotspots to appear. Astradot allows instant, dynamic, re-partitioning of your topics. Scale producers and consumers hetrogenously as your workload changes. Need 10 producers and 100 consumers, you can do that! This results in no performance hotspots ever!
No network costs
At scale, the cross-az network costs dominate all other costs of running Kafka. Astradot requires no cross-az replication since it writes directly to S3. Since traffic to S3 is free, Astradot has no network costs at all.
Durability and Availability of S3
Astradot writes not just your data to S3 but also a vast majority of related metadata. Thus you are getting all the availability and durability guarantees of S3.
Performance
Kafka Producers are the only place where latency is higher than regular Kafka. Its not significantly higher though since produce requests are coalesced. Talk to us to get more details on this.
Kafka Consumer performance is similar to regular Kafka, as Astradot is able to use pre-fetching of data.
Astradot is written in Rust, so you get the performance of native code with no GC.
100% Kafka compatibility
Fully api-compatible with Kafka, including all security features of Kafka.
Enterprise Ready
Multi-region
AWS PrivateLink support
Cluster level RBAC
SSO/SAML support
SOC2 Type2 audited
About us
We have been operating in stealth mode out of San Francisco. We are backed by True Ventures, with participation from GGV Capital, Uncorrelated Ventures and other prominent investors.