A unified benchmark shows learned indexes beat fence pointers on memory-latency tradeoff in LSM-trees, with position boundary and SSTable granularity as the key tuning knobs.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DB 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Evaluating Learned Indexes in LSM-tree Systems: Benchmarks,Insights and Design Choices
A unified benchmark shows learned indexes beat fence pointers on memory-latency tradeoff in LSM-trees, with position boundary and SSTable granularity as the key tuning knobs.