MeshTest: End-to-End Testing for Service Mesh Traffic Management
summary
MeshTest automatically generates end-to-end service mesh configs using a novel Service Flow Exploration to enumerate resource interactions, then translates configs into low-level control flow graphs and uses a Service Mesh Oracle (SMT-based) to synthesize and check representative requests. Implementation targets Istio and Linkerd, finds 23 new bugs (19 confirmed, 10 fixed), generates ~2,500 configurations/sec, and runs ~29 requests per input on average. Building per-mesh interpreters costs about two person‑weeks and the framework achieves 100% coverage on pairwise resource interactions.
AI-written summary · about
chapters (9)
- 0:00MeshTest introduction
- 1:42Service mesh testing needs
- 3:09Unit test limitations
- 4:13Config generation challenges
- 6:40MeshTest automation
- 9:08Control flow testing oracle
- 11:39Evaluation results
- 12:43Bug found demo
- 13:24Conclusion
related talks
28mService Meshes: At What Cost?
49mSolving the Service Mesh Adopter's Dilemma
43mWTF Do We Need a Service Mesh?
34mService Mess to Service Mesh
11mWelcome & Opening Remarks
44mTesting Service Mesh configs & k8s manifests
39mService Mesh Specifications and Why They Matter in Your Deployment
20mThe Crazy Service Mesh Ecosystem