Distributed search

How this question breaks down

1. 1Inverted indexCore21 minFlip documents into a map from each word to the documents that contain it, so a query reads a few short lists instead of every document.
2. 2Scatter-gatherCore27 minSplit the index by document across shards, send each query to one copy of every shard, and merge their top results. Shards buy size; replicas buy throughput.
3. 3RankingCore24 minScore every match by how well its words fit the query and how good it is on its own, and find the top ten without scoring everything carefully.

Work through the parts in order and you’ve covered the full design. Each part is one topic.