Several municipal Lambdas need shared read access to a large GIS tile cache with POSIX file semantics that exceed practical /tmp size. What should they mount?
Select an answer to reveal the explanation.
Short Explanation
When many functions need the same big map-tile library with real file paths, EFS is the shared network drive—/tmp is too small and per-package copies are absurd. Glacier-per-request would melt your latency budget.
Full Explanation
Lambda can mount Amazon EFS to access a shared filesystem when object storage alone is a poor fit and /tmp capacity is insufficient. That enables POSIX semantics across concurrent functions for large read-mostly caches. Embedding huge caches in deployment packages, forcing Glacier downloads per invoke, or misusing DynamoDB as a POSIX filesystem are poor fits.