3.12.x → 4.x Migration tool is available to all Hazelcast Enterprise customers. To get the tool or learn further details, please reach out to Hazelcast Support.
Hazelcast IMDG 3.9
- Extended Support for IMDG 3.9 will end on 23 April 2021.
Hazelcast IMDG 3.12
- Standard Support for IMDG 3.12 will end on 9 April 2021.
- Extended Support will continue for an additional 18 months.
For more information see the full list of Version Support Windows.
Latest News / Releases
Hazelcast IMDG 3.12.12
- Improved the handling of failures for queries to eliminate the client application's indefinite hanging when the user code deployment feature is used.
Hazelcast JET 4.4
- Introduced the beta version of Jet's SQL interface. You can now log into Jet from the command line and issue queries against the data sources you specify. See the documentation here.
- Introduced the Unified File Connector API to create sources and sinks from file-like resources. See the documentation here.
- Implemented the connector for Amazon Kinesis Data Streams which is a durable, scalable real-time data streaming service native to the AWS environment. See its tutorial here.
- Introduced prevention for event reordering; by default Jet reorders data for performance, now you can disable this to get strict event order where you need it. See the documentation here.
- Improved the naming of source and sink stages across different connectors, bringing them all in line with the same convention.
- Added the pipeline.isEmpty() method that tells whether the pipeline contains any stage.
- Added a configuration property that disables the enforcement of the exact IMDG dependency version, which allows adding IMDG quick fixes to the existing Jet release.
- Introduced a slim Jet package; you can now either download the full Jet package with all the extensions included or the slim one with no Jet extensions giving you the freedom of downloading the desired extensions separately.
- Improved the behavior of submitting jobs; this process now waits for the job to start and informs this with a message.
- Improved the Kafka source performance by fine-tuning various timeouts.
Hazelcast Cloud 2.7
- Introduced Social Sign-in allowing you to login to Hazelcast Cloud with your Google or GitHub account.
- Added Hazelcast IMDG 3.12.10 and 4.1 versions to Hazelcast Cloud.
Hazelcast Cloud 2.6.3
- Improved the payment approval handling.
Hazelcast .NET Client 4.0
New Features & Enhancements
- Improved the API to fully embrace the .NET asynchronous patterns using async and await calls.
- Improved the low-level networking to rely on high-performance .NET libraries that are also used in critical products such as the Kestrel web server.
- Redesigned the configuration around the Microsoft.Extensions.Configuration abstractions while allowing it to be compatible with the traditional .NET configurations.
Redesigned the logging to rely on the Microsoft.Extensions.Logging abstractions while allowing to be compatible with the traditional .NET logging.
See here for details on the solved issues and enhancements that have been involved in this version release.
Hazelcast C++ Client 4.0
- Implemented the new Hazelcast Open Binary Client Protocol version 2 for this client.
Upgraded the client code to use the C++11 features.
Improved the API methods so that they return boost::future which allows async programming.
- Introduced the new hazelcast::new_client API to obtain and start a new client.
Added support of cmake install and cmake find_package(hazelcast) for projects using Hazelcast.
- Replaced the AP concurrency primitives with the CP ones (with respect to the CAP principle) that includes Atomic Long, Atomic Reference and Semaphore structures.
Introduced boost::optional for return types in the API which makes it possible to return nullable objects.
- Added the support of backup acknowledgements (aka boomerang backups) so that the client does not wait for the sync backups to complete on the member anymore.
- Added the support of mixed type so that you do not need to provide a specific object type when instantiating a Hazelcast structure such as a map.
Made Boost the only external dependency for the client.
- Added the support of using asio for SSL and non-SSL sockets.
Fully switched to the lock based atomic implementations of the std::atomic standard library.
- Introduced fail-fast for the serialization implementations that cannot be resolved during the compilation.
Replaced group name element with cluster name in the client configuration (ClientConfig), which also has replaced the GroupConfig.
Introduced new logger configuration and listener APIs.
Renamed the camelCase API as snake_case.
Removed the getConnectionAttemptLimit and setAttemptPeriod configurations and replaced them with the exponential backoff configuration connection_retry_config in the connection strategy configuration.
Hazelcast Python Client 4.0
- Applied the changes that have come with Hazelcast IMDG 4.0 to the client protocol and architecture, including the ownerless client feature.
- Simplified the design to track member leaves and joins to the cluster.
- Implemented CP concurrency primitives (with respect to CAP principle) including AtomicLong, AtomicReference and Semaphore.
- Added the support of backup acknowledgements (aka boomerang backups).
- Added the support of paging predicates.
- Simplified the programmatic configuration to use the keyword arguments.
- Aligned the docstrings with the Google Python Style Guide.
- Improved the logging configuration API to follow the practices recommended by the standard library documentation.
- Improved the client performance for the non-blocking mode.
- Moved the predicates from the serialization module and improved their documentation.
- Improved the performance of string serialization.
- Added the support of native bytearray and Java UUID types for the default serialization.
- Implemented the set_ttl method for the maps.
- Moved the API and detailed documentation to a new platform, i.e., hazelcast.readthedocs.io.
- Implemented the missing ObjectDataInput#skip_bytes method.
- Changed the time resolution unit for all APIs to seconds.
- Improved the responsiveness of the client in case of the reconnection scenarios.
- Improved the enum-like configurations so that they can be customized using string literals.
- Reformatted the source code of the client using
- Removed the AP-based concurrency primitives due to the implementation of CP-based ones.
- Renamed various configuration options due to the simplifications in the programmatic configuration.
- Removed the custom handlers from the loggers due to the improvements in the logging configuration API.
- Renamed various predicates that has been resulted from changing their belonging module (hazelcast.serialization.predicate → hazelcast.predicate)
- Removed the private APIs from the public ones and their documentation.
- Updated the listener APIs to return Future[str] and Future[bool] (instead of str and bool) due to since you can now add or remove listeners in a non-blocking way.
- Renamed the hazelcast.exception module as hazelcast.errors.
Hazelcast Node.js Client 4.1
- Added the support of blue/green deployment and disaster recovery for the client.
- Added a new discovery mechanism that allows the smart client to decide whether it needs to use public addresses for communication with members.
- Added the support of IMap.setTtl() and IAtomicLong.getAndDecrement() operations.
- Introduced the maxIdle parameter for the map operations that insert and modify map entries.
- Added the capability of reporting statistics to Hazelcast Management Center.
- Eliminated the random connection delay when a port is not specified to connect to the members
- Increased the default cluster connection timeout from 20 seconds to 2 minutes to provide more reliable behavior in Kubernetes and similar environments.
- Increased the default automated pipelining threshold from 8KB to 64KB.
Hazelcast Node.js Client 4.0.2
Hazelcast Tomcat Session Manager 2.2
- Added automatic configuration for Spring Boot.
- Added the support of YAML configuration.
Best wishes from the Hazelcast Team!