ITG Engineering Guidelines Page 119 of 378
ITG Trunk 2.0 ISDN Signaling Link (ISL) Description, Installation and Operation
The technician can decide to keep costs down, an d accept a te mporary "Fa ir"
QoS level for a select ed route. I n tha t cas e, ha ving m ade a ca lcul ated t rade -off
in quality, you need to carefully monitor the QoS level, reset expectations
with the end users, and be receptive to user feedback.
Fine-tune Networ k QoS
Topics presented in this section d eal with issues t hat will impact t he QoS of
ITG traffic. They are informative for understanding how to fine-tune a
network to improve its QoS, but are not directly involve d as a part of network
engineering procedure. These are advanced topics to help a technician fine
tune the network to improve QoS, but they are not a part of the required
procedure for initial ITG network engineering.
Further network anal
y
sis
This section describes actions that could be taken to investigate the sources
of delay and error in the intranet. This and the next section discuss several
strategies for reducing one-way delay and pac ke t loss . The key strat egies are :
• Reduce link delay
• Reduce hop count
• Adjust jitter buffer size
• Implement IP QoS mechanisms
Components of delay
End-to-end delay is contributed by many delay components; the major
components of delay are described as follows.
Propa
g
ation dela
y
Propagation delay is affected by the mileage and medium of links traversed.
Within an average size country, the one-way propagation delay over
terrestrial lines is under 18 ms; within the U.S. the propagation delay from
coast-to-coast is under 40 ms. To estimate the propagation delay of long-h aul
and trans-oceanic circuits use the rule-of-thumb of 1 ms per 100 terrestrial
miles.
If a circuit goes through a satellite system, estimate each hop between earth
stations to contribute 260 ms to the propagation delay.