Chaos Search seems to be doing this architecture already and according to the podcast episode [1], it uses a highly optimized storage layout. Never used it, so

Chaos Search seems to be doing this architecture already and according to the podcast episode [1], it uses a highly optimized storage layout. Never used it, so would be interested if somebody could comment on it. [1] https://www.dataengineeringpodcast.com/chaos-search-with-pete-cheslock-and-thomas-hazel-episode-47/ https://www.dataengineeringpodcast.com/chaos-search-with-pet...