Skip to content

Read vs. Write Asymmetry

Read and write operations have fundamentally different characteristics in most storage systems: reads can be cached and parallelized freely, while writes require coordination and durability guarantees.

Tension Addressed

Sequential vs. Random Access Optimization

Recognizable As

Read replicas, write-through vs write-back caches, CQRS patterns, separate read/write paths in databases, "read-heavy" vs "write-heavy" workload characterization

Canonical Sources

Gray & Reuter, "Transaction Processing" (1992); Hellerstein, Stonebraker & Hamilton, "Architecture of a Database System" (2007)

Transfers To

Database scaling strategies, distributed system design, cache invalidation, replication architectures

This asymmetry drives architectural decisions like separating read replicas from write masters, and influences cache coherency strategies.