Kubernetes and Service Mesh Upgrade Automobile Company’s IT Infrastructure
summary
Large Chinese auto manufacturer moves from VM + self-built DNS/ELB + per-VM nginx to Kubernetes + service mesh in production, replacing manual service registration with an automated control plane and transparent sidecar proxies. Built features include weighted canary 30/70 independent of instance count, mTLS automated cert rotation and L7/L4 traffic matching, language-agnostic telemetry, gradual traffic cutover per service (vehicle-charging tested), and measurable ops reductions and lower ID investment.
AI-written summary · about
chapters (19)
- 0:00Start
- 2:35Current architecture
- 3:55Abstract architecture(Current) Integrated platform ServiceMeshCon
- 4:43Abstract architecture (Target): Cloud Native infrastructure ServiceMeshCon
- 5:48Service discovery and load balance: Current ofütión
- 6:24Service discovery and load balance: Current sõíútíón
- 7:09Service discovery and load balance: Compare ServiceMeshCon
- 8:32Canary release: Current solution
- 9:08Canary release: Target solution(2)
- 11:52Security: Current solution
- 12:35Security: Target solution
- 13:32Security: Compare
- 14:24Observability: Current solution
- 14:54Observability: Target solution
- 15:31Observability: Compare
- 16:29Target Cloud native architecture
- 18:20Comparison summary
- 20:15Migration solution
- 22:11Failover between containers and VMs
related talks
39mDo you need a Service Mesh?
47mDeploying Service Mesh in Production
53mDo you need a Service Mesh?
41mService Mesh beyond single clusters
36mOnly Slightly Bent: Uber’s Kubernetes Migration Journey for Microservices
58mUsing Service Meshes and Kubernetes to Solve Service-to-Service Communications...
21mProduction-Ready gRPC: How Service Meshes Simplify Security, Routing, and Observability
34mService Mess to Service Mesh
