Replication:

We use Passive Replication. The replication manager would take care of starting a new middleware server. This would be the eventual strategy. As of now for the next demo we are not bringing the replication manager in. We would detect an error and switch to an already running middleware. This is a kind of Active Replication.

Fault Detection:

This includes timeout exceptions for transactions between remote components, such as the client to middleware servers. The faults at all the three tiers are caught.

Failover:

As of now the failover mechanism is shifting over to another middleware server. A replication manager

Client Identification:

The client sends his login as identification. It is concatenated with the time stamp and logged upon every transaction.