On the previous pages of our
series on software caches in Java application servers,
we focused on an example application using caches as a simple,
stand-alone, temporary data store.
The following is an (incomplete) list of advanced features that have
not been covered on this site.
| Ehcache | Hazelcast | Infinispan | |
|---|---|---|---|
| Off-Heap Memory: Cached data can be held in a part of the memory that is not managed by the garbage collector to improve vertical scalability. | yes | yes | soon |
| Persistent Caches: Cached data can be stored on hard disk without integrating a 3rd party database. | yes | no | no |
| Full-Text Search: Cached data can be made searchable, e.g. via Lucene indexes. | yes | no | yes |
| WAN Replication: Clusters can be replicated across geographically separated sites. | yes | yes | ? |
| GUI Tools for Maintenance: Tools visualizing maintenance data about the cache deployment. | yes | yes | no |
| JMX Management: Expose maintenance data about the caches via JMX. | yes | yes | yes |
| Messaging and Processing: Cluster can be used as a distributed messaging and processing platform. | no | yes | no |
The summary recapitulates some of the results.