The Path to Real-Time with SingleStoreDB Self-Managed 5 at Strata+Hadoop San Jose

The Path to Real-Time with SingleStoreDB Self-Managed 5 at Strata+Hadoop San Jose

At Strata+Hadoop World, we engaged with the brightest minds in data management, machine learning, and analytics. CEO and co-founder, Eric Frenkiel announced the release of SingleStoreDB Self-Managed 5 and showcased a path for real-time processing to the Strata audience during the day one keynote. Later, he shared the stage with Kellogg in a tutorial session on predictive analytics.

single-store-db-self-managed-5-now-generally-availableSingleStoreDB Self-Managed 5 Now Generally Available

Eric announced the general availability release of SingleStoreDB Self-Managed 5, delivering breakthrough performance on database, data warehouse, and streaming workloads.

SingleStoreDB Self-Managed 5 features:

LLVM Code Generation Architecture

  • LLVM code generation for SQL queries, featuring a new SQL to LLVM to machine code compiler built into the the engine
  • Completely revamped code generation architecture to greatly improve query compilation speed

Query Optimization Improvements

  • Updated query optimizer uses more collected statistics and histograms to choose better query plans
  • Improvements to hash join selection, distributed and local join optimization, and better selection of bushy joins
  • Elimination of unnecessary tables, views, and filters
  • Intelligent selection of distributed group by execution plans

Columnstore Performance Improvements

  • Operations on compressed data
  • Batch scanning to read data and apply filters in batches for faster scan performance
  • Prefetching for improved query execution when the table does not fit in memory

SQL Features

  • New EXPLAIN shows an updated, informative description of query execution plans
  • Support for temporary tables
  • Several windows functions and the OVER clause

View the full SingleStoreDB Self-Managed 5 release notes ⇒

driving-the-on-demand-economy-with-predictive-analytics-keynoteDriving the On-Demand Economy with Predictive Analytics – Keynote

Eric explains how a trinity of real-time technologies—Kafka, Spark, SingleStore— enables predictive applications and analytics for enterprise companies.

dash-forward-from-descriptive-to-predictive-analytics-with-apache-sparkDash forward: From Descriptive to Predictive Analytics with Apache Spark

JR Cahill, Senior Solutions Architect at the Kellogg Company, joined Eric for a tutorial session outlining Kellogg’s approach to advanced analytics with SingleStore. JR shared how Kellogg moved from overnight to intraday analytics, integrating directly with Tableau. Follow SingleStore on Twitter to see the full recording as soon as it’s available.

Follow SingleStore on Twitter ⇒


Share