The stack provides automatic connection migration, i.e. the connections are moved automatically by the stack. The migration is triggered by the failure of a server. If a connection had a backup server, then it is moved to this server, and the connection may be recovered by the distributed application there.

Synchronization between the primary and backup TCP/IP Stacks can be achieved in two ways:

• With active synchronization, where the primary always keeps its backups up to date. This means there is an interchange of information between them every time the state of the connection changes.
• With passive synchronization. The primary does not keep its backups updated. When there is a failure, the primary recovers the state using the information kept by the client and the distributed application. The state from the client is recovered using legal TCP exchanges, and does not require any change in its stack.