“…1) rely on the routing protocol to re-route affected sessions and simply restart the AC process each time a session is re-routed [15], [16], [19], [28], [29], [30], [31], [33], [32], [34], [36], [37], [39], [40], [41], [42], [43]; 2) notify source nodes which then decrease the rate of, or pause packet sending for affected sessions. This at least avoids the development of congestion, but these sessions must then be re-admitted anyway [11], [25], [38], [24], [45]; 3) or attempt fast local route repair to limit the QoS assurance violation time and the development of congestion [18], [27].…”