Class ReplicatedHashMap<K extends Serializable,V extends Serializable>

java.lang.Object
java.util.AbstractMap<K,V>
org.jgroups.blocks.ReplicatedHashMap<K,V>
All Implemented Interfaces:
ConcurrentMap<K,V>, Map<K,V>, ReplicatedMap<K,V>, ExtendedMembershipListener, ExtendedMessageListener, ExtendedReceiver, MembershipListener, MessageListener, Receiver

public class ReplicatedHashMap<K extends Serializable,V extends Serializable> extends AbstractMap<K,V> implements ConcurrentMap<K,V>, ExtendedReceiver, ReplicatedMap<K,V>
Subclass of a ConcurrentHashMap with replication of the contents across a cluster. Any change to the hashmap (clear(), put(), remove() etc) will transparently be propagated to all replicas in the group. All read-only methods will always access the local replica.

Keys and values added to the hashmap must be serializable, the reason being that they will be sent across the network to all replicas of the group. Having said this, it is now for example possible to add RMI remote objects to the hashtable as they are derived from java.rmi.server.RemoteObject which in turn is serializable. This allows to lookup shared distributed objects by their name and invoke methods on them, regardless of one's onw location. A ReplicatedHashMap thus allows to implement a distributed naming service in just a couple of lines.

An instance of this class will contact an existing member of the group to fetch its initial state.

This class combines both

invalid reference
org.jgroups.blocks.ReplicatedHashtable
(asynchronous replication) and
invalid reference
org.jgroups.blocks.DistributedHashtable
(synchronous replication) into one class
Author:
Bela Ban