Ritesh Yadav
Research Notes
Kubernetes Networking
A request to a Kubernetes Service crosses several layers before it reaches an application. These notes explain those layers using small examples, then look at what changes as a cluster grows.
Start here
- Why Cilium replaces kube-proxy: follow a payments request, see how Service and endpoint state grows with more microservices, and compare iptables rules with eBPF lookups.
Related reading
- LLM serving on Kubernetes: application gateways, inference routing, and model servers above the Pod network.
About the author
Ritesh Yadav works as an AI/ML Engineer. He writes independent research notes on ML performance, infrastructure, and systems, covering CUDA, low-latency inference, generative AI, distributed training, Kubernetes, and LLMOps.