Skip to content
Switch branches/tags


Failed to load latest commit information.
Latest commit message
Commit time


Slack GitHub javadoc Docker pulls Total Alerts Code Quality: Java Quality Gate Status

Hazelcast is a distributed in-memory data store and computation platform that's fault tolerant and easy to scale up or down.

As an in-memory data store, Hazelcast gives you faster access to your data by storing it in memory. With more accessible data, you can leverage Hazelcast to process huge amounts of real-time events or static datasets with consistently low latency.

To help you take advantage of all these features, Hazelcast comes with the following built-in data structures:

  • a distributed, partitioned and queryable in-memory key-value store implementation, called IMap
  • additional data structures and simple messaging constructs such as Set, MultiMap, Queue, Topic
  • cluster-wide unique ID generator, called FlakeIdGenerator
  • a distributed, CRDT based counter, called PNCounter
  • a cardinality estimator based on HyperLogLog.

Also, Hazelcast includes a production-ready Raft implementation which allows implementation of linearizable constructs such as:

  • a distributed and reentrant lock implementation, called FencedLock
  • primitives for distributed computing such as AtomicLong, AtomicReference and CountDownLatch.

Hazelcast data structures are in-memory, highly optimized and offer very low latencies. For a single get or put operation on an IMap, you can typically expect a round-trip-time of under 100 microseconds.

Additionally, Hazelcast provides a distributed batch and stream processing engine named Jet. It provides a Java API to build stream and batch processing applications through the use of a dataflow programming model. You can use it to process large volumes of real-time events or huge batches of static datasets. To give a sense of scale, a single node of Hazelcast has been proven to aggregate 10 million events per second with latency under 10 milliseconds.

Jet enables Hazelcast to import/export data from/to a very wide variety of data sources such as Apache Kafka, Kinesis, Local Files (Text, Avro, JSON), JDBC, JMS, Elasticsearch, Apache Hadoop (Azure Data Lake, S3, GCS) and much more.

It's very simple to form a cluster with Hazelcast, you can easily do it on your computer by just starting several instances. The instances will discover each other and form a cluster. There aren't any dependencies on any external systems.

Hazelcast automatically replicates data across the cluster, and you are able to seamlessly tolerate failures and add additional capacity to the cluster when needed.

Hazelcast comes with clients in the following programming languages:

Hazelcast also has first-class support for running on different cloud providers such as AWS, GCP and Azure as well as on Kubernetes.


You can download Hazelcast from Once you have downloaded, you can start the Hazelcast instance using the script bin/

Get Started

Hazelcast allows you to interact with a cluster using a simple API, for example you can use the Hazelcast Java Client to connect to a running cluster and perform operations on it:

HazelcastInstance hz = HazelcastClient.newHazelcastClient();
IMap<String, String> map = hz.getMap("my-distributed-map");
map.put("key", "value");
String current = map.get("key");
map.putIfAbsent("somekey", "somevalue");
map.replace("key", "value", "newvalue");

You only need to add a single JAR as a dependency:


For more information, see the Getting Started Guide


See the reference manual for in-depth documentation about Hazelcast features.

Code Samples

See Hazelcast Code Samples

Get Help

You can use the following channels for getting help with Hazelcast:

Using Snapshot Releases

Maven snippet:

    <name>Sonatype Snapshot Repository</name>

Building From Source

Building Hazelcast requires JDK 1.8. Pull the latest source from the repository and use Maven install (or package) to build:

$ git pull origin master
$ mvn clean install

Take into account that the default build executes thousands of tests which may take a considerable amount of time. Additionally, there is a quick build activated by setting the -Dquick system property that skips tests, checkstyle validation, javadoc and source plugins and does not build extensions and distribution modules.


Hazelcast has 3 testing profiles:

  • Default: Type mvn test to run quick/integration tests (those can be run in parallel without using network).
  • Slow Tests: Type mvn test -P slow-test to run tests that are either slow or cannot be run in parallel.
  • All Tests: Type mvn test -P all-tests to run all tests serially using network.


Hazelcast uses static code analysis tools to check if a Pull Request is ready for merge. Run the following commands locally to check if your contribution is Checkstyle compatible.

mvn clean validate


Source code in this repository is covered by one of two licenses:

  1. Apache License 2.0
  2. Hazelcast Community License

The default license throughout the repository is Apache License 2.0 unless the header specifies another license.


Thanks to YourKit for supporting open source software by providing us a free license for their Java profiler.


Copyright (c) 2008-2021, Hazelcast, Inc. All Rights Reserved.

Visit for more info.