CONCEPT Cited by 1 source

Time Travel (Data State)

Definition

Time travel in the data-movement context refers to the ability to process the schema, cluster topology, and data as a cohesive unit at a specific historical point in time — enabling auditing, debugging, disaster recovery, and reproduction of past data states.

Why Casspactor couldn't do it

The legacy Casspactor assembled backup metadata from multiple services at query time. If the cluster topology or keyspace schema changed between the backup and the retrieval, Casspactor could not reconstruct the prior state (Source: sources/2026-06-19-netflix-the-evolution-of-cassandra-data-movement-at-netflix).

How the new stack enables it

The Cassandra Analytics Wrapper reads all necessary context (SSTable data, schema, topology metadata) directly from the S3 backup at a specific snapshot point, treating the backup as a self-contained, immutable artifact.

Related concepts

• systems/apache-iceberg — provides table-level time travel via snapshot versioning.
• concepts/snapshot-isolation — the transactional analogue at the database layer.

Seen in

• sources/2026-06-19-netflix-the-evolution-of-cassandra-data-movement-at-netflix — critical feature of the new data-movement stack.