interviewed – Paul Groom
summary
Claims that Hadoop cannot handle high-frequency BI workloads directly and that a scale-out in-memory platform fixes that by sitting between BI tools and Hadoop. Platform uses standard SQL/ODBC/JDBC, drives hundreds of CPU cores across nodes, and can hold tens to hundreds of terabytes of RAM so BI queries meet strict latency and concurrency SLAs. Demonstrated approach targets reducing operation onslaught from Hadoop into low-latency interactive analytics and is presented as working in production-class environments.
AI-written summary · about
related talks
47mDistributed high-load BI engine for Google Data Studio and Microsoft Power BI
59mHadoop Essentials for the SQL Server Professional
07mIn-Hadoop Analytics: Bringing analytics to big data
45mSolving Real & Big (Data) Problems Using Hadoop
45mFrom Big Data to Big Information
50mFast Analytics On Big Data
39mBig Data Systems Performance: The Little Shop of Horrors
49mBuilding a 100‑Petabyte Analytics Warehouse