It should be noted that even for a "small" environment we recommend 32gb of RAM for the Primary Hub Server and it should be a dedicated machine for the task in most cases. It is not unusual for some of the Java probes (discovery_server, wasp, etc) to be allocated up to 4gb each. A java.exe using 1gb is normal and in fact you may have more than one.
The hub.exe itself should not be using 11gb however and this could be an indication of a queue that is backed up. You can check in the Infrastructure Manager GUI under "Status" and look at the queue status - if any queue has a large number of messages listed in the "Queued" column this can be an indicator that the queue is not flowing well, and the messages in the queue can sometimes consume a lot of memory when this occurs.