Page 140 of 378 ITG Engineering Guidelines
553-3001-202 Standard 1.00 April 2000
gives time for implementation processes to follow through. The planning
thresholds can be set 5% to 15% below the QoS objec ti ves, de pending o n th e
implementation lag time.
Intranet QoS monitoring
To monitor one-way delay and pac k et loss statistics, you mus t install a delay
and route monitoring tool, such as ping and traceroute, on the T- L AN
of each ITG site. Each delay monitoring tool will be running continuously,
injecting probe packets to each ITG site about every minute. The amount of
load generated by this is not considered important. At the end of the month,
the hours with the highest one-way delay are noted; within those hours, the
packet loss and standard deviation statistics can be computed.
(See “Measure intranet QoS” on page114 for information about
implementation of the ping hosts and the use of scripting.)
(See “Obtain QoS measurement tools” on page 118 for information about
where to obtain other more specialized delay and route monitoring tools.)
At the end of the month, the technician can analyze each site’s QoS
information. Table23 provides a sample.
Declines in QoS can be observed through the comparison of QoS between last
period and current period. If a route does not meet your QoS objective, you
must take immediate action to improve the route’s performance .
Table 23
QoS monitoring
Site pair One-way delay
Mean+σ (ms) Packet loss
Mean+σ (%) QoS
Last
period
Current
period
Last
period
Current
period
Last
period
Current
period Objective
Santa Clara/
Richardson 135 166 1 2 Excellent Good Excellent
Santa Clara/
Ottawa 210 155 3 1 Good Excellent Excellent
Etc.