SCIENCE

SGI has announced support and record benchmarks for open-source VoltDB SQL database version 1.1. SGI Rackable scale-out cluster delivered over three million transactions per second with near linear scalability running the VoltDB "Voter" application. VoltDB's "Voter" simulates the high-frequency and large transaction volumes experienced by Internet- and cloud-based financial trading, social networking, ad-serving, online gaming and telecommunications companies.

VoltDB is the next-generation open-source SQL database management system (DBMS) built for online transaction processing (OLTP) applications. Its performance advantage stems from an in-memory, distributed database partitioning concept optimized to run on today's memory-rich servers with multicore CPUs.

"We are pleased with the excellent scaling and performance results of VoltDB running on our hardware and system software stack," said Christian Tanasescu, vice president of software engineering at SGI. "SGI's Rackable cluster running VoltDB 1.1 directly translates into a clear price and performance advantage for Internet-based companies having to process massive amounts of transactions at any given moment."

VoltDB automatically "shards" the database into data partitions and distributes each partition -- along with an SQL engine for processing -- to a different CPU core in the SGI Rackable cluster. Each high-frequency Intel Xeon processor with 8MB L3 cache executes its VoltDB database transaction queue at in-memory speeds. VoltDB automatically maintains data consistency across all nodes in the cluster and replicates data throughout the SGI Rackable cluster to provide high availability year-round.

"Organizations no longer have to struggle with database scalability limitations, or give up SQL and data consistency to overcome them," said Scott Jarr, president and CEO of VoltDB. "VoltDB and the SGI Rackable cluster are a natural match for developers and customers who require the best database price and performance using scale-out hardware."