Hazelcast

Hazelcast is an open-source in-memory data grid and stream processing platform, written in Java. Founded in 2008 by Fuad Malikov, Talip Ozturk, and Greg Luck, Hazelcast has grown into a distributed computing platform used by thousands of organizations including Capital One, UBS, Cisco, and T-Mobile. With over 6,200 stars as of 2026, Hazelcast provides distributed data structures, distributed computing, and stream processing capabilities on a single unified platform. Key features include: distributed in-memory data store (distributed implementations of Java collections including IMap for concurrent distributed map, IQueue, ISet, IList, ITopic, and MultiMap with partition-based data distribution across cluster nodes for near-infinite horizontal scalability), distributed computing (ExecutorService for executing tasks on specific cluster members, EntryProcessor for atomic in-place map entry processing without serialization overhead, and callable/future abstractions for distributed task execution), event-driven stream processing (Hazelcast Jet, a stream processing engine built on Hazelcast providing high-throughput, low-latency data pipelines with windowing, aggregation, joining, and mapping operations, capable of processing millions of events per second), SQL support (querying distributed maps and Kafka topics via standard SQL queries), near cache (client-side caching of frequently accessed entries for sub-millisecond read latency), HD memory (off-heap memory storage for high-density caching without GC pressure), WAN replication (cross-datacenter replication for disaster recovery and geo-distribution), partition service (data partitioning across nodes with backup partitions for high availability), entry listeners and event listeners (notifying on map modifications), query support (SQL-like predicates, custom predicates, and indexing for fast lookups), and integration with Kafka, JCache (JSR-107), Spring, and JPA.