Consider the graph to the right, which represents the
computer network. Each line represents an established connection. Therefore, the server
C is connected directly to
A, which is also connected to
B and
D. If a disruption in the connection between
C and
A occurs, the connection may be terminated as a result. This can occur either by a
socket producing an error, or by excessive
lag in which the far server
A anticipates this case (which is called a timeout). When the connection between
A and
C is severed, users who were connected to other servers that are no longer reachable on the network appear to quit. For example, if user
Sara is connected to server
A, user
Bob is connected to server
B, and user
Joe is connected to
C, and
C splits, or disconnects, from
A, it will appear to
Joe as if
Sara and
Bob both quit (disconnected from the network), and it will appear to both
Sara and
Bob that
Joe quit. However,
Joe can still talk to anyone who is connected to the same server (in this case server
C). This happens because the servers to which they are connected are informed of the change in the network status, and update their local information accordingly to display the change. Later, server
C may relink (reconnect) to a server (or servers) on the network and the users who appeared to have quit will rejoin; the process of sending this updated information to all servers on the network is called a
netburst (or
sync). ==Security issues==