I'm not sure exactly how this will work but from a discussion on the Infinispan mailing list discussion (subject: Infinispan URL format):
Question:
Once you are connected to a remote (Infinispan) database, does the
application simply use the java.util.Map api to put/get any application
get values? Or are puts not allowed to use application classes? I'm
trying to better understand how the marshaling works, since the remote
Infinispan database probably wouldn't have access to the application
classloader (unless it does, which I'd like to also understand).
Answer:
Scott, the application receives Infinspan's CacheManager and/or Cache
APIs just like in the Mongo case, one receives the Mongo specific
objects.
As far as the objects you can put in the cache: the ideal situation is
that you use a protobuf schema and the client side will marshall things
as protobuf and send these protobuf structure to the server. The server
then does not need to have the client classes in its classpath.
I'm not sure exactly how this will work but from a discussion on the Infinispan mailing list discussion (subject: Infinispan URL format):
Question:
Answer: