Handle cache-update conflicts in distributed services
Company: eBay
Role: Software Engineer
Category: System Design
Difficulty: hard
Interview Round: Onsite
Your service uses a cache (e.g., Redis) in front of a relational database, and multiple application servers may update the same key concurrently. Design how to prevent and resolve cache–database update conflicts and race conditions: discuss write ordering and atomicity, versioning/ETags or compare-and-set, distributed locks, TTLs and invalidation vs write-through/write-back, idempotent updates, and recovery after partial failures. Specify APIs, data model, consistency guarantees, and trade-offs.
Overview: This question evaluates a candidate's competency in distributed systems and data consistency, focusing on cache–database consistency, concurrent write handling, versioning/CAS, distributed locking, cache invalidation strategies, idempotency, and failure recovery in high-throughput services.
Read the full eBay Software Engineer interview experience this question came from