Memory‑vs‑Disk Trade‑Off: LiipImagine Bundle Cache Backend for Sulu Asset Management

Goal

Determine the optimal cache strategy for on‑the‑fly image transformations in Sulu’s Asset Management module.

Constraints & Uncertainty

• Containerized deployments often have strict RAM limits, yet image resizing demands low latency.
• The LiipImagineBundle supports an in‑memory array cache (default) and a file‑based cache; the Sulu docs do not prescribe which to use.
• Cache invalidation on asset replacement is only governed by TTL, risking stale images until the TTL expires.

Unresolved Decision

Should the cache be kept in memory to prioritize speed for small assets, or switched to a file‑based cache to conserve RAM in constrained environments?

Questions

1. What is the typical memory footprint of the in‑memory cache for a standard set of resized images in a production Sulu instance?
2. How does the file‑based cache affect response times under concurrent image resize requests compared to the array cache?
3. Does Sulu provide a mechanism to force immediate cache invalidation when an asset is updated, or is it solely TTL‑driven?