Keynote: State of gRPC
summary
gRPC celebrates ten years and showcases concrete adoption: Netflix uses it since 2018, Spotify since 2019, and LinkedIn migrates about 50,000 endpoints migrated onto gRPC, proving large-scale viability. It uses Protocol Buffers and HTTP/2 for smaller messages, low latency and high throughput, and supports large payloads for ML training and inference (TensorFlow, Triton). Google Cloud highlights a service mesh without sidecars that pushes load balancing and retries into the app, and a public preview of gRPC Rust is arriving soon.
AI-written summary · about
chapters (8)
- 0:00gRPC introduction
- 1:26gRPC adoption success
- 2:57Why gRPC excels
- 4:41gRPC in Cloud & mesh
- 5:45Rust support future
- 6:50gRPC in AI/ML
- 8:56New AI protocols
- 9:40Closing remarks
related talks
18mgRPC and AI
32mEnhancing the Performance Testing Process for gRPC Model Inferencing at Scale
12m10 Years of CNCF and Cloud Native's Next Decade
19mBeyond JSON-RPC: Scaling Model Context Protocols With gRPC in the Py...
37mEfficient IoT with Protocol Buffers and gRPC
16mKeynote: LLM-Aware Load Balancing in Kubernetes: A New Era of Efficiency
30mSpotify Harnesses Ray for Next-Gen AI Infrastructure
34mKeynote: Welcome + Opening Remarks