In-Memory databases (Hyper, HANA) often favor Adaptive Radix Trees (ART). Why?

1. A They utilize significantly fewer disk seeks than standard B-Trees when operating on exceptionally large datasets that exceed the system's RAM capacity.
2. B They provide a more robust framework for managing ACID-compliant transactions in a distributed environment with high network latency.
3. C They optimize for CPU Cache Locality by utilizing wide nodes and byte-chunked keys to minimize the frequency of expensive L3 cache misses.
4. D They are significantly easier to persist to durable storage devices during sudden system failures due to their linear, log-structured memory layout.