Big Data, Big Performance

on YouTube ↗09m

" summary

HyperTable is a C++ BigTable clone with a Thrift API and production deployments at Baidu and Rediff Mail (100 million users) storing metadata and metrics. Benchmarks on a 16‑node cluster with 5 TB show HyperTable roughly 2x faster than HBase for random reads and writes and HBase suffers GC collapse at small value sizes. Features include a query cache, atomic counters with no random IO, regex filters using Google re, namespaces and 1 billion cells inserted in 45 minutes via secondary indexes.

AI-written summary · about

" chapters (5)

1. 0:00Start
2. 5:12Random Write Test
3. 7:09Regular Expression Filtering
4. 7:54Secondary Indexes
5. 8:35Namespaces

" related talks

• 01h 00mHyper Speed: When Big Data BloomsScott Helme · NDC Oslo 2021
• 01h 00mHyper Speed: When Big Data BloomsScott Helme · NDC Security 2022
• 01h 06mYahoo!, Big Data, and Microsoft BI: Bigger and Better TogetherDenny Lee · PASS Business Analytics Conference 2013
• 26mColumnar Storage: Redefining Data Management for the Modern EraZoe Steinkamp · Voxxed Days Thessaloniki 2024
• 38mCharacterizing the performance of Big Data processing platformsManish Singh · Data By the Bay 2016
• 42mBillion-Scale Triumph: Innovating at the Edge of Cloud DataBalázs Szántó · Craft Conference 2025
• 01h 03mThe Future of Apache Hive and Hadoop 2.0: Bigger, Faster, StrongerCarter Shanklin · PASS Business Analytics Conference 2013
• 53mBig Data: from mammoth to elephantRoman Nikitchenko · JavaDay Lviv 2015