Visualizers
Learn by moving things
Interactive explainers for the ideas behind retrieval, embeddings, and agents. A formula tells you what is true; dragging a vector shows you why. Each one also appears inside the article it belongs to, and runs entirely in your browser, so drag, poke, and break it.
Least response time
From: How a Load Balancer Spreads the Load →Send to the server answering fastest right now, so traffic steers away from a struggling server even before it fails a health check. Needs constant latency measurement.
Least response time: send to the server answering fastest right now. It naturally steers away from a slow or struggling server.
Power of two choices
From: How a Load Balancer Spreads the Load →The modern default: do not check every server, pick two at random and send to the less busy one. Almost the evenness of checking everyone, for almost none of the cost, and it avoids the herd.
Power of two choices: pick two servers at random, then send to the less busy of the two. Almost as good as checking them all, at a fraction of the cost.
Source-IP hash
From: How a Load Balancer Sticks →Hash the client's IP to pick a server, so a given client always lands on the same one. No cookie, works at layer 4. Watch out for many clients behind one address.
Hash the client's IP to pick a server, so a given client always lands on the same one. No cookie needed, and it works at layer 4.