LLM-D 1.0, AN OPEN SOURCE STACK FOR DISTRIBUTED INFERENCE

Summary from the TLDR IT newsletter

An open, Kubernetes-native serving stack with cache-aware routing, multi-tenant traffic control, tiered KV cache, and high availability ready for production on Google Cloud.