Data Plane · Streaming & Stream Processing

Redpanda

Kafka API-compatible streaming data platform written in C++.

Overview

Redpanda is a Kafka-compatible streaming platform rewritten from scratch. Producers and consumers use the Kafka API unchanged, so it is a drop-in replacement at the client level, but nothing inside is Kafka's code.

The design is the pitch. It is written in C++ with a thread-per-core model, using the Seastar library, to get more out of the same hardware, with Raft for replication.

It ships as a single binary with the schema registry, HTTP proxy and broker built in, and no external dependencies at all, meaning neither ZooKeeper nor KRaft nor a JVM.

The vendor's claims are specific: ten times lower average latencies than Kafka on identical hardware, and three to six times fewer compute resources.

Be precise about the licence, because "open-source Kafka alternative" would be wrong. The community edition is under the Business Source License, which forbids offering Redpanda as a commercial streaming service and converts each merge to Apache 2.0 four years later.

The enterprise feature list is long enough to matter when comparing: tiered storage, role-based access control, audit logging, Iceberg topics, cloud topics, Kerberos and OIDC authentication, FIPS compliance and whole-cluster restore are all licensed features.

Tiered storage offloads log segments to object storage, and the newer cloud topics put topic data directly there, enabled by default when object storage is configured in the current release.

The lakehouse feature is Iceberg topics, storing topic data in Iceberg format so Snowflake, Databricks, ClickHouse and Redshift can read it, with either in-bucket cataloguing or an Iceberg REST endpoint.

The AI direction is an agentic data plane, with Model Context Protocol integration, more than 300 connectors including AI and GPU services, and bring-your-own-cloud deployment.

Pricing is the weak spot for comparison. Redpanda publishes essentially no unit rates, only trial credits, and its price estimator is a contact form.

Key features and capabilities

The same headings are used for every streaming & stream processing entry, so two tools can be read side by side.

What it does
  • A Kafka-compatible partitioned log, with topics, partitions and consumer groups over the Kafka API
  • Written in C++ on a thread-per-core model, using Raft for replication
  • A single binary including the broker, schema registry and HTTP proxy, with no ZooKeeper, KRaft or JVM
  • Cloud topics, an enterprise feature, place topic data directly in object storage
Delivery and ordering guarantees
  • Kafka protocol semantics, so per-partition ordering, idempotent producers and transactions all carry over
  • Replication is Raft-based rather than Kafka's in-sync replica mechanism
  • Deferred upgrade finalisation with guaranteed rollback in the current release
  • Client quota metrics are reported per entity
Scale and retention
  • Serverless limits, 100 MB/s in, 300 MB/s out, 5,000 logical partitions, unlimited retention and storage
  • Bring your own cloud goes to 2 GB/s write, with a 99.99% availability commitment against 99.9% on serverless
  • Tiered storage, an enterprise feature, offloads segments to object storage for long retention
  • Tiered storage v2 is in beta and adds compaction for object-stored topics
Processing and transformation
  • No Kafka Streams equivalent in the broker; Redpanda Connect handles transformation
  • Redpanda SQL is listed as a product on the vendor site
  • Kafka-ecosystem processors such as Flink, Spark and Kafka Streams work over the Kafka API
  • An agentic AI service with Model Context Protocol integration for agent workloads
Security and governance
  • Authentication through SASL/SCRAM, mutual TLS with principal mapping, and cloud IAM roles
  • Authorisation through ACLs, with role-based and group-based access control licensed as enterprise features
  • TLS and mutual TLS encryption, with a configurable FIPS-compliant mode
  • Audit logging, Kerberos, OIDC and schema registry authorisation are all enterprise-only
  • Community-edition users should check the feature list carefully before assuming parity
Connectors and ecosystem
  • Kafka API compatibility gives access to existing Kafka clients and connectors
  • Redpanda Connect ships more than 300 connectors, including AI and GPU services
  • Iceberg topics with in-bucket cataloguing or an Iceberg REST catalog endpoint
  • A built-in schema registry and HTTP proxy in the same binary
Where it runs and what it costs
  • Self-managed on containers, laptops, x86 and ARM hardware, cloud, bare metal or edge
  • An official Kubernetes operator with GitOps compatibility, plus Helm, on GKE, EKS, AKS and bare metal
  • Redpanda Cloud in serverless, dedicated and bring-your-own-cloud forms
  • Community edition is source-available rather than open source, so redistribution terms need checking

Pricing

Price on requestQuote only; free trial credits published

Redpanda publishes no unit rates. Serverless is described as pay as you go or annual commitment, bring your own cloud as an annual commitment, and dedicated capacity carries no published price at all; the estimator page is a sales contact form. The only hard figures are trial credits, $100 for 30 days on AWS serverless, or $300 through the AWS and Google Cloud marketplaces. Self-managed enterprise pricing is not published, though new clusters get a 30-day trial licence, and the community edition is free to download under the Business Source License.

Vendor pricing page →

Demos and videos

About Redpanda Data

Redpanda Data was founded in 2019 by Alex Gallego, is based in San Francisco and runs a fully remote team across more than 17 countries. It is privately held and venture-funded, raising a Series A led by Lightspeed with GV in 2021, $100m Series C in June 2023 taking total capital to $165m, and a $100m Series D led by GV in April 2025 at a $1bn valuation, announced with its agentic AI platform. Licensing is two-tier: the community edition under the Business Source License, converting to Apache 2.0 after four years, and the enterprise edition under the Redpanda Community License with a licence key.

Founded 2019 · San Francisco, California · redpanda.com

Other streaming & stream processing tools

Amazon Kinesis

Data Plane · Streaming & Stream Processing

AWS services for collecting and processing real-time streaming data.

  • Cloud service

Apache Flink

Data Plane · Streaming & Stream Processing

Stateful stream processing framework for real-time computation over unbounded data.

  • Open source

Apache Kafka

Data Plane · Streaming & Stream Processing

Distributed event streaming platform for high-throughput data pipelines and streaming applications.

  • Open source

Azure Event Hubs

Data Plane · Streaming & Stream Processing

Managed event ingestion service on Azure, with a Kafka-compatible endpoint.

  • Cloud service

Confluent

Data Plane · Streaming & Stream Processing

Data streaming platform built around Apache Kafka, available as a cloud service or self-managed.

  • Commercial

Drafted with AI assistance and checked against the vendor’s own documentation.