Techment — Site Header

What is Avro

Definition

Apache Avro is a compact, schema-based binary serialization format that stores the schema alongside the data and supports rich schema evolution, making it a popular choice for streaming systems and message payloads where producers and consumers must evolve independently.

« Back to Glossary Index

Embeds or references the schema so data is always self-describing on read
Handles schema evolution gracefully, letting fields be added or changed safely
Produces compact binary payloads ideal for high-throughput streaming
Integrates tightly with schema registries for compatibility enforcement

Real World Example

A Kafka-based event platform serializes messages with Avro and registers each schema, so producers can add optional fields over time while existing consumers keep deserializing older and newer messages without errors.

FAQs

What makes Avro good for streaming?

Its compact binary format, self-describing schemas, and strong schema-evolution support make it ideal for evolving event streams.

How does Avro handle schema changes?

It uses writer and reader schemas with compatibility rules, allowing fields to be added or removed without breaking consumers.

How does Avro differ from Parquet?

Avro is row-oriented and optimised for streaming and write-heavy use, while Parquet is columnar and optimised for analytical reads.