Case Study: ICE Mortgage Technology achieves high-performance, scalable in-memory caching and reduced TCO with Hazelcast

Ellie Mae Chooses Hazelcast for its Stability, Performance, Durability, and Scalability

ICE Mortgage Technology (formerly Ellie Mae) runs the Encompass SaaS platform for mortgage origination and faced persistent performance and scalability problems: disk-bound database access caused high latencies and a homegrown in-memory cache prevented horizontal scaling and lacked high availability, threatening SLAs and rapid business growth. ICE evaluated vendors and turned to Hazelcast, selecting its in-memory data grid and High-Density Memory Store to meet strict latency, eviction, consistency and .NET client requirements.

Hazelcast replaced the homegrown cache with a clustered Hazelcast deployment and High-Density Memory Store, enabling ~100 GB of in-memory data, explicit locking and backups for high availability, and extensive Simulator-driven testing. The Hazelcast solution dramatically reduced DB load and delivered predictable low latency (SLA of 5 ms met; benchmark latencies <2 ms), sustained throughput (average >31,500 TPS across 8 nodes, peaking at 45,000 TPS), low heap usage and linear scalability—providing ICE Mortgage Technology the performance, availability and lower TCO it required.

ICE Mortgage Technology