How and Why Netflix Built a Real-Time Distributed Graph: Part 3 — Querying the graph with gRPC execution API - Devoured

Back to Devoured - August 10, 2026

DEVOURED

How and Why Netflix Built a Real-Time Distributed Graph: Part 3 — Querying the graph with gRPC execution API

Data infrastructurebackend Netflix

Netflix manages 150 billion edges in its real-time distributed graph by optimizing query execution with gRPC and level-by-level parallel traversal.

What: The graph query layer achieves P99 latency under 150ms for three-hop traversals by utilizing 16 to 24 async threads per instance and aggressive caching tuned to data volatility.

Decoder

gRPC: A high-performance remote procedure call framework developed by Google that uses Protocol Buffers for structured data serialization.
TTL (Time to Live): A setting that defines the lifespan of data in a cache before it is invalidated or refreshed.

Original article

Netflix's RDG query layer handles 8 billion nodes and 150 billion edges using level-by-level parallel traversal, 16 to 24 async threads per instance, and cache TTLs matched to data volatility that achieve 70 to 80% hit rates on stable entities. The result is that even three-hop traversals complete under 150 ms at P99.

Explore all 68 articles in today's digest