https://devopstales.github.io/tags/llm-hosting/ Recent content in Llm-Hosting on DevOpsTales Hugo -- gohugo.io en-US Mon, 14 Sep 2026 00:00:00 +0000 https://devopstales.github.io/ai/ollama-to-vllm-migration/ Mon, 14 Sep 2026 00:00:00 +0000 https://devopstales.github.io/ai/ollama-to-vllm-migration/ Ollama is the easiest way to run a local model, until the experiment becomes a shared service. Requests queue, latency wobbles, prefixes recompute, and one GPU stops being enough. vLLM answers exactly those problems — but migration is a trade, not an upgrade: simplicity for scheduling, memory control, parallelism, and production operations.