Cold kiln archives are rarely reread; hot daily logs are scanned every training run; a third pile is a sparse wide table. How should the team map those piles, and what must they not do?
Select an answer to reveal the explanation.
Short Explanation
Think of cold kiln archives, hot daily logs, and a sparse wide table. Map each pile by cost, performance, and structure. Do not train directly from Glacier.
Full Explanation
Initial storage decisions come from cost, performance, and structure: S3 Standard for hot scans, Glacier for cold archives, a file-system store when POSIX I/O is required. Never train directly from Glacier.