Infinispan 7.2.0 Final 发布，此版本主要有以下更新：
Bug 修复：
•[ISPN-2813] - DummyTransaction do not commit is last resource transaction is read-only
•[ISPN-3361] - Tests from org.infinispan.rest.IntegrationTest failing
•[ISPN-3725] - Recovery does not work with versioning
•[ISPN-4167] - DummyBaseTransactionManager doesn't clean up when rollback fails
•[ISPN-4322] - AS Remote Client Module Integration Tests module fails with IBM JDK
•[ISPN-4442] - Package remoting jars for the embedded cli client
•[ISPN-4450] - RemoteStore uses wrong classloader with rawValues
•[ISPN-4546] - Possible stale lock when the primary owner leaves during rebalance
•[ISPN-4986] - Error in CDI User Guide
•[ISPN-5019] - After coordinator change, cache topologies should be installed in parallel
•[ISPN-5106] - Deadlock on GlobalComponentRegistry when starting a cluster
•[ISPN-5139] - MissingFormatArgumentException in trace logging in ClusteredCacheLoaderInterceptor
•[ISPN-5141] - SingleFileStore.process() should use sequential access
•[ISPN-5158] - Transaction rolled back but returns successful response
组件升级：
•[ISPN-5325] - Upgrade to protostream-3.0.1.Final
•[ISPN-5395] - Upgrade Infinispan Server to WildFly 8.2.0.Final
性能增强：
•[ISPN-4797] - Use getAll for retrieving query results
•[ISPN-5085] - Convert REST server to using RESTEasy's Netty container integration
•[ISPN-5147] - Creation/configuration of caches should not require a server restart
•[ISPN-5231] - CacheNotifierImpl should wire the listener and converter with dependecies from ComponentRegistry
•[ISPN-5243] - Configuration attribute holders
•[ISPN-5251] - Eviction configuration must be enforced in Near Caches
•[ISPN-5264] - Optimize PutAll operations (library mode)
•[ISPN-5265] - Optimize Getall operations over Hot Rod
•[ISPN-5266] - Optimize PutAll operations (Hot Rod)
•[ISPN-5283] - Hot Rod client does not update the cluster view for the client after a complete shutdown of all servers
•[ISPN-5306] - Eviction should allow for long size configuration
•[ISPN-5320] - Use new bounded CHMv8 for Hot Rod client near caches
•[ISPN-5331] - Enable node statistics with global jmx enabled flag
•[ISPN-5349] - Use DSL based filters with local and clustered event listeners
•[ISPN-5350] - Use DSL based filters with Hot Rod event listeners
•[ISPN-5370] - Make clear() non-transactional and lock free
•[ISPN-5388] - Optimize redundant predicates (non-indexed query)
•[ISPN-5390] - QueryCache should use the internal cache registry
•[ISPN-5391] - Use @MetaInfServices in infinispan-directory-provider module
•[ISPN-5401] - Align server and embedded exposed management properties
•[ISPN-5410] - Stronger isolation between modules needing Infinispan Core and those needing Lucene
•[ISPN-5435] - Add annotation attribute to allow for listener to specify whether it receives pre, post or all events
Infinispan 是个开源的数据网格平台。它公开了一个简单的数据结构（一个Cache）来存储对象。虽然可以在本地模式下运行Infinspan，但其真正的价值在于分布 式，在这种模式下，Infinispan可以将集群缓存起来并公开大容量的堆内存。这可比简单的复制强大的多，因为它会为每个结点分配固定数量的副本----服 务器故障的一种恢复手段----同时还提升了可伸缩性，这是由于存储每个结点所需的工作量是与集群大小息息相关的。
Infinispan 提供了一种简单的机制来利用大容量的堆内存。如果对每个结点维护一个拷贝，假如集群当中有100个结点，每个结点分配2GB的堆内存， 那么网格中的任何实例都能使用多达100GB的空间，这可都是内存，显然速度会非常快。同时Infinispan还兼容于JTA，这样它就能很好地处理事务 了。我们还有一个超级强大的异步API，它可以保证同步的网络调用以及异步调用的并行性及可伸缩性。比方说：Future f = cache.putAsync(k, v) 可以阻塞线程，再调用f.get()可以让网络调用继续进行或是忽略掉f。更为重要的是，线程还可以做别的事情，这一点非常有用。然后再回来通过调用 f.get()来检查该网络调用是否能继续进行。可以将其看作是NIO与传统的阻塞性IO之间的关系。
Infinispan公开了一个 CacheStore接口和几个高性能的实现，包括JDBC CacheStores、基于文件系统的CacheStores以及Amazon S3 CacheStores等等。CacheStores可用作“温启动（warm starts）”或是确保网格中的数据在重启后依然可用，同时在内存耗尽时还能将数据写到磁盘上。
主要特点：
大量的堆体
极高的可扩展性
快速轻量级核心
不仅仅支持Java(PHP,Python,Ruby,C…)
支持Compute Grids
管理是关键：当你在grid上运行几百个服务时，实现管理是必须的
软件详情：https://issues.jboss.org/secure/ReleaseNote.jspa?projectId=12310799&version=12325808
下载地址：http://infinispan.org/download/
来自:开源中国社区
