daslab.seas.harvard.edu lab Splaying LSM-Trees T. Lively, L. Schroeder, C. Mendizábal Advisor: Stratos Idreos The Problem: LSM-Trees Have Suboptimal Read Performance Recently updated keys are the cheapest to access …but these keys are not necessarily most likely to be read next Copy frequently accessed keys to top Use ’get’ as a trigger for possible ‘put’ …and over time the tree is reorganized for better read performance Compared with the state-of-the-art solution of adding a cache: …and can offer adaptivity and better performance across workloads Splaying dominates cache in the absence of Bloom Filters The Solution: Splaying
1
Embed
E-Tree: An Ever -Evolving Tree for Evolving Workloads Splaying …daslab.seas.harvard.edu/doc/undergrad-poster/SIGMOD-UG-splaying-l… · By tracking access patterns, subtrees dynamically
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.