· Security
Why APISIX For AI Gateway Workloads
A pragmatic analysis of APISIX for LLM traffic, with the SafeLLM integration perspective.
Tag: performance
Semantic Caching: What It Actually Saves (and What It Does Not)
An 80% cache hit rate does not mean 80% cost reduction. Here is an honest breakdown of what caching saves, what it costs to operate, and when you should not cache at all.