Skip to main content
Data governance keeps the data on your streams intelligible: producers and consumers agree on the structure of a message, and that agreement is enforced rather than assumed. On StreamNative Cloud, schema management is the foundation of that. A Pulsar cluster can run two schema registries:
  • The Pulsar Schema Registry, built into every broker. Schemas are stored per topic and managed with Pulsar tooling.
  • The Kafka Schema Registry, available when the Kafka protocol is enabled on the cluster. It’s compatible with the Confluent Schema Registry API, stores schemas per subject, and is what Kafka clients expect.
A schema registered in one registry is invisible to the other. There’s no synchronization between them. Decide which registry a topic uses before you build against it, and make sure every producer and consumer on that topic agrees.
A native Kafka cluster has only the Kafka Schema Registry. Everything below about choosing between registries applies to Pulsar clusters. See Kafka Schema Registry for the Kafka cluster documentation.

​
Compare the options

This leaves you three ways to attach schemas to data on a Pulsar cluster.

​
Pick this if

Use the Pulsar Schema Registry when your producers and consumers are all Pulsar clients and you have no need to share schemas with Kafka applications. It requires no extra service, no extra endpoint, and no extra credentials—the schema travels with the topic. It also supports schema types the Kafka registry has no equivalent for, including KeyValue and the primitive types. Use the Kafka Schema Registry when your applications are Kafka clients. Standard Confluent serializers and deserializers work unchanged, so an application written against another Confluent-compatible registry can point at StreamNative Cloud without code changes. Use the Kafka Schema Registry from Pulsar clients when both protocols read the same topic—for example, a Kafka application produces the data and a Pulsar application consumes it. The External Schemas library gives you a Pulsar Schema implementation backed by the Kafka serializers, so you keep the familiar Pulsar Producer and Consumer APIs while schemas resolve from the Kafka Schema Registry.

​
Next steps

Pulsar Schema Registry

Concepts, schema types, compatibility, and management.

Kafka Schema Registry

Enable it on a Pulsar cluster and connect your clients.

Use Kafka schemas from Pulsar clients

Produce and consume Kafka-schematized data with the Pulsar APIs.

Replicate schemas

Set up active-standby Schema Registry replication.