High Performance Service Meshes with gRPC Proxyless
summary
gRPC proxyless embeds the data plane in applications to remove sidecar overhead and claims measurable gains: benchmarks show higher QPS, lower memory use, and faster tail latency versus sidecars. It implements XDS for control-plane compatibility, supports mTLS, client-side circuit breaking, retries, outlier detection, fault injection, and pluggable load balancers. OpenTelemetry metrics are available and Google Cloud offers a Gateway API + Traffic Director workflow and a Cloud Service Mesh preview; a rate‑limit RLQS integration lands first for Java.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 3:32Service mesh evolution
- 7:47Proxyless gRPC benefits
- 9:11Key features
- 11:40Service mesh planes
- 13:27Gateway APIs
- 15:36Cloud service mesh updates
- 16:21Q&A
related talks
29mgRPC load balancing and Service Mesh
24mgRPC Proxyless Service Mesh
29mThe Security Evolution of GRPC Services in the Mesh
35mService Mesh use cases for Telco and Edge
31mThe Future of Service Meshes: Sidecarless or eBPF?
30mTo Mesh or not to Mesh
28mProxy Service: A New Network Traffic Abstraction in Kubernetes
35mgRPC Load Balancing on Kubernetes - Jan Tattermusch, Google