VoltDB has a long heritage that dates back to the hallowed halls of the most groundbreaking engineering institutions on the planet. The original H-Store academic research paper was a collaboration of senior computer scientists — led by Dr. Michael Stonebraker — from MIT, Yale University, and Brown University. H-Store’s objective was to design a natively scalable, fault-tolerant, transactional database ready for the data revolution of the 21st century. These visionaries accurately predicted the onslaught of data that would be driven by the increasingly on-demand nature of an online, high-velocity world, and they realized that a new kind of database would be required for companies to respond and thrive. This groundbreaking work was commercialized by, and is now available from VoltDB.

RDBMS Processing GraphH-store was the seminal work that broke the mold that traditional relational database management systems (RDBMSs) have forced for years – a “one size fits all” approach. Traditional systems were bound by decades old architectures (see diagram). This wasn’t a problem when data volume and velocity and complexity were manageable. Today, however, data has become so fast that general purpose RDBMSs simply cannot keep up. And any attempt to get them to do so ends up creating multiple points of failure and drives costs that are prohibitively high.

By focusing on the problem of high throughput operations and in-memory analytics, VoltDB has made forward-thinking architectural choices that achieve multiple orders of magnitude and more operational throughput than legacy systems on a per-machine basis. But single-machine performance is just one dimension of performance.

Designed for shared-nothing clusters of commodity hardware, where each node has many cores, VoltDB embraces parallelism natively. Where traditional systems use shared locks and latches to leverage multi-core systems, VoltDB achieves concurrency through scheduling. It treats each core in a multi-node cluster as a pipeline to fill with work. In this way, linear scalability is achieved as machines are added.

In the end, VoltDB is an in-memory database designed to handle the most challenging workloads and operations at tremendous speeds with limitless scalability and zero data loss. Its revolutionary, scale-out architecture allows VoltDB users to build once-thought-impossible applications that leverage inexpensive, virtualized computing infrastructures, including popular cloud service platforms.