Case Study: SBI Securities achieves 20× lower latency, 3–4× higher throughput and reduced costs with Pivotal GemFire (Pivotal)

SBI Securities Achieving Significant ROI with Improved Performance and Scalability

SBI Securities, Japan’s largest online broker with over 2.6 million accounts and global trading operations, faced a mission-critical performance problem: its three-tier, RDBMS-driven trading system hit a database bottleneck as accounts and transaction volumes grew. An initial fix—duplicating and partitioning systems—reduced performance pain but increased costs, complexity, and operational risk, so the company needed a new, scalable data-tier approach that preserved reliability and customer satisfaction.

Working with NRI and Hitachi, SBI implemented Pivotal GemFire, an in-memory distributed data grid, replacing the duplicated infrastructure and transforming the application layer. The deployment cut servers from 400 to 150 and shrank the data center footprint fivefold, while delivering 20x better latency (0.05s vs 1s), 3–4x higher throughput, and the ability to handle twice the prior transaction volume—reducing costs and operational complexity while providing elastic, highly available scale.

SBI Securities