implement a spliterator like interface for parallel traversal of local and replicated cache

The backport of ConcurrentHashMapV8 comes with an interesting interface, Spliterator and it's 3 implementations, Key/Value/EntryIterator.

As java 7 is now being more widely adopted, it could be interesting to offer to take advantage of the java ForkJoinPool for the parallel traversal of local and replicated caches where the Infinispan map reduce framework is maybe less suited for.

At term, it could be used to speed up the search in a cache without the need to index it.