21.6
NETWORK MANAGEMENT FROM A PSTN PERSPECTIVE
551
is usually less than the total switching capacity. When the load is at or below engineered
capacity, the switch handles calls in an efficient and reliable manner. However, when the
load increases beyond the engineered capacity, delays can occur internal to the switch. This
resulting internal congestion can spread, causing connected systems to wait for start-dial
indications. This can also cause internal congestion in connected switching systems.
At the onset of overload, also known as circuit shortage, the dominant cause for
customer blockage is the failure to find an idle circuit. Circuit blockage alone limits the
number of extra calls that can be completed but does not cause a significant loss in call-
carrying capacity of the network below its maximum. As the overload persists and the
network enters a congested state, regeneration-calling pressure changes customer blockage
from circuit shortage to switching delays.
Switching delays cause timeout conditions during call setup and occur when switches
become severely overloaded. Timeouts are designed into switches to release common-
control components after excessively long delay periods and provide the customer with
a signal indicating call failure. Switching congestion timeouts with short holding-time
attempts on circuit groups replace normal holding-time calls. Switching delays spread
rapidly throughout the network.
A trunk-group overload usually occurs during general or focused overloads and/or
atypical busy hours. Some of the overload causes not discussed above are facility outages,
inadequate trunk provisioning, and routing errors. The results of trunk-group overload can
be essentially the same as those previously discussed for general overloads. However, the
adverse effects are usually confined to the particular trunk group or the apex area formed
by the trunk group and those groups that alternate-route to the overloaded trunk-group.
Trunk-group overload problems can often be minimized or handled completely by the use
of temporary NTM reroute controls until a more permanent solution can be provided.
21.6.5
Network Traffic Management Controls
21.6.5.1
Circuit-Switched Network Controls
. There are two broad categories of
NTM controls:
z
Protective Controls. These controls remove traffic from the network during overload
conditions. This traffic is usually removed as close as possible to its origin, thus
making more of the network available to other traffic with a higher probability
of completion.
z
Expansive Controls. These controls reroute traffic from routes experiencing over-
flows or failures to other parts of the network that are lightly loaded with traffic
because of noncoincident trunk and switching system busy hours.
Implementation of either type of control can be accomplished on a manual or automatic
basis. For example, manual controls are activated by network traffic managers, and auto-
matic controls are activated by network components. In some switches, these controls are
implemented on a planned control-response basis that is preprogrammed into the switch.
In other systems, controls are available on a flexible basis, whereby any control can be
assigned to any trunk group on a real-time basis.
The availability of any specific control, its allowable control percentages, and the
method of operation can vary with the specific type of switch. In many instances, these net-
work controls can be activated with variable percentages of traffic affected (for example,
25%, 50%, 75%, and 100%) to fine-tune the control to match the magnitude of the