TABLE OF CONTENTS Many media-stimulated overload events are known about in advance, and so network management overload controls (such as call gapping) could be applied to try to limit the calling rate to one that will just keep terminating lines busy, or prevent switches getting congested. However, as sections 7.4.1 and 7.4.2 illustrate, the complexity of the network makes setting the overload controls to throttle call streams by the right amount a very difficult task. It is even harder to cope with the many unannounced events that occur. We conclude that overload controls, to be effective, need to be automatic and embedded in the network resources in order to react sufficiently rapidly to sudden surges of demand or changes in network capacity.
For any specific network context, the number of nodes causing a resource to be overloaded can vary widely from one event to another, for example:
in national circuit-switched networks, the number of neighbouring switches overloading a target switch can range from a few (fewer than 10) up to many (of the order of 100);
in intelligent networks, SCPs can receive service requests from several hundred service-switching points (SSPs) (see Chapter 8 for details) in certain national overload scenarios all of the SSPs can cause the SCP to be overloaded, while for a regional overload, the number of sources of the overload could be very small;
in ATM networks, the originating node selects the entire end-to-end path across a peer...
