Why did you choose a dictionary as the cache storage?

A dictionary provides average O(1) lookup, insertion and removal by key. It fits a cache lookup API, but does not provide TTL, eviction or thread safety by itself.

Choose it when the main operation is finding a value by a stable identifier. Add expiry metadata for TTL and an explicit capacity policy; an LRU cache usually also needs an ordered structure.

Synchronize shared mutable access, and define whether returned reference values may be modified. Average constant time is not a worst-case guarantee, and hashing cost depends on the key. Explain these trade-offs instead of claiming that a dictionary is automatically a complete cache.