English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

All data is not equally popular. Often, some portion of data is morefrequently accessed than the rest, which causes a skew in popularity of thedata items. Adapting to this skew can improve performance, and this topic hasbeen studied extensively in the past for disk-based settings. In this work, weconsider an in-memory data structure, namely hash table, and show how one canleverage the skew in popularity for higher performance. Hashing is alow-latency operation, sensitive to the effects of caching, branch prediction,and code complexity among other factors. These factors make learningin-the-loop especially challenging as the overhead of performing any additionaloperations can be significant. In this paper, we propose VIP hashing, a fullyonline hash table method, that uses lightweight mechanisms for learning theskew in popularity and adapting the hash table layout. These mechanisms arenon-blocking, and their overhead is controlled by sensing changes in thepopularity distribution to dynamically switch-on/off the learning mechanism asneeded. We tested VIP hashing against a variety of workloads generated byWiscer, a homegrown hashing measurement tool, and find that it improvesperformance in the presence of skew (22% increase in fetch operation throughputfor a hash table with one million keys under low skew, 77% increase undermedium skew) while being robust to insert and delete operations, and changingpopularity distribution of keys. We find that VIP hashing reduces theend-to-end execution time of TPC-H query 9, which is the most expensive TPC-Hquery, by 20% under medium skew.

VIP Hashing -- Adapting to Skew in Popularity of Data on the Fly (extended version)