1 of 1 people found this helpful
Not sure about this, but I'd imagine that for user-defined keys, you'd use the grouping API to collocate the data: https://docs.jboss.org/author/display/ISPN/The+Grouping+API
Grouping API is really an adavntage to collocate the keys on specific nodes. But need some help to understand the article in detail.
Our understanding is that, the hash code of the key is not imporrant and only the hash code of the group is know to all the nodes. So whenever a request for a key lands in the cache, the group is determined at run-time using some algorithms and the key is added to a node where the group belongs.
But what is not clear is how do we know to which node a group belongs ? Is there a way to configure the groups to specific nodes. This could be beneficial in knowing the group and its node in advance so that we can redirect the requests to only those knows which are owners of the key.