Is the underlying cache implementation aware of (and "participating in")
ongoing JTA transactions?
Regions which report that they are transaction-aware are considered
"synchronous", in that we assume we can immediately (i.e. synchronously)
write the changes to the cache and that the cache will properly manage
application of the written changes within the bounds of ongoing JTA
transactions. Conversely, regions reporting false are considered
"asynchronous", where it is assumed that changes must be manually
delayed by Hibernate until we are certain that the current transaction
is successful (i.e. maintaining READ_COMMITTED isolation).