Clustering is enabled and initialized by installing a valid clustering license. In this case, the caches are provided and managed by Hazelcast. On the other hand, when clustering is not enabled, caching is provided by Google’s ConcurrentLinkedHashMap library, which is a high performance version of java.util.LinkedHashMap for use as a software cache. For details see, https://code.google.com/p/concurrentlinkedhashmap/.
In both clustered and non-clustered cases, caching is configured and used in the same unified way.