Dar, S., Franklin, M. J., Jónsson, B. Þ., Srivastava, D. and Tan, M. (1996). Semantic data caching and replacement. In Proc. VLDB 1996, pp. 330–341.
Key takeaways
Read in full (NOTE-904). A client cache described by the selection predicates of past queries can answer a new selection query partly from cache (the probe query, Q ∧ V) and fetch only the rest from the server (the remainder query, Q ∧ ¬V), and can evict by semantic region rather than by page or tuple. In one simulator, one client and one server, read-only selection queries on a single 10,000-tuple relation, this matched page caching on a clustered index, beat it heavily when clustering was bad, answered over 60% of small queries on an unindexed attribute with no server contact, and let a value function based on semantic distance beat LRU on a skewed workload. No replacement rule won everywhere: on regular cyclic paths MRU was best. Everything is measured, nothing is proved; updates, joins, and the CPU cost of reasoning about cache contents were not modelled.
Why this is here
CLAIM-303: a cache keyed by the predicates of past queries answers a new query partly from cache and sends only the remainder to the server. Partial reuse depends on stored answers carrying a description in the same language as the questions, which a filed argument lacks. Simulation only.