“Performance is everything when you’re researching global financial markets because the amount of data to be analyzed grows every day,” said Alex Davies, head of production engineering at Jump Trading Group. “In order for quants to be productive, they need instant access to massive amounts of data. VAST Data’s cost-effective, elegant all-flash architecture was the clear choice for our read-heavy research workloads.”
Many of the world’s leading thinkers, researchers and quants agree that VAST Data’s ability to provide fast access to data is a critical requirement for HPC environments--especially for algorithmic trading where an ever-increasing amount of data has to quickly be made available to researchers. Announced this week at the 2021 Supercomputing Conference, VAST Data was selected as the award winner for HPCwire’s Readers’ Choice awards for best use of HPC in financial services. The award recognized Jump’s implementation of VAST Data’s Universal Storage platform.
How VAST Data Accelerates Trading
The financial trading market is fiercely competitive. Firms constantly struggle to model markets accurately, which requires copious amounts of data just to train the machine learning models. As the amount of data grows, legacy systems based on hard disk drives (HDDs) just aren’t able to keep up. Jump removed spinning drives from its research pipelines entirely to improve performance.
“When you have new financial markets such as crypto appearing out of nowhere, you suddenly have an enormous amount of data that you need to store and access quickly,” continued Davies. “VAST solves this because its all-flash infrastructure allows us to quickly and independently scale our systems to store petabytes of data.”
In addition to being more performant than HDD-based storage solutions, the unique compression algorithms embedded into VAST’s Universal Storage software allow the company to provide customers with a more efficient platform that requires less power. It accomplishes this through VAST’s similarity-based data reduction algorithms which allow customers to further reduce data, thus increasing available storage capacity. All-flash systems also utilize less parts than HDD so there are fewer points of failure. The result is higher performance and reliability at a lower total cost of ownership.
“Jump Trading Group is truly an innovative financial research company that has devoted significant resources to engineering that keeps them at the very top of the game,” said VAST Data Co-Founder Jeff Denworth. “They’re competing on their ability to utilize AI and machine learning which depend on fast and reliable data pipelines and we’re happy to play a small part in their success.”
VAST Data completely changes the economics of flash storage with its Universal Storage platform. By consolidating multiple storage tiers into a single global namespace, data access accelerates and application performance improves enabling financial institutions to quickly process market data and AI data sets at any scale. Visit VAST’s website to learn more about Universal Storage for algorithmic trading.