
Engineering guidelines 23
P0911590 Issue 02 Enterprise Edge 2.0 IP Telephony Configuration Guide
Determining WAN link resources
For most installations, IP telephony traffic is routed over WAN links within the
intranet. WAN links are the most expensive recurring expenses in the network and
they often are the source of capacity problems in the network. Different from LAN
bandwidth, which is almost free and easily installed, WAN links, especially inter-
LATA and international links require time to receive financial approval, provision
and upgrade. For these reasons, it is important to determine the state of WAN links
in the intranet before installing the IP telephony.
Each voice conversation, (G.729, Annex B codec, 20 ms payload) uses 12 kbit/s of
bandwidth for each link that moves across in the intranet; a DS0 can support below
5 simultaneous telephone conversations.
Link utilization
The starting point of this evaluation is to get a current topology map and link
utilization report of the intranet. A visual inspection of the topology map indicates
which WAN links are expected to deliver IP telephony traffic. Also use the
Traceroute tool (see Measuring Intranet QoS on page 27).
The next step is to find out the current utilization of those links. Note the reporting
window that appears in the link utilization report. For example, the link utilization
can be an average of a week, a day, or one hour. To be consistent with the
dimensioning considerations (see Enterprise Edge VoIP Gateway bandwidth
engineering on page 17), get the peak utilization of the trunk. Also, because WAN
links are full-duplex that data services show asymmetric traffic behavior, get the
utilization of the link representing traffic flowing in the heavier direction.
The third step is to determine the available spare capacity. Enterprise Edge intranets
are subject to capacity planning controls that ensure that capacity use remains
below some determined utilization level. For example a planning control can state
that the utilization of a 56 kbit/s link during the peak hour must not exceed 50%; for
a T1 link, the threshold is higher, for example at 85%. The carrying capacity of the
56 kbit/s link can be 28 kbit/s, and for the T1 1.3056 Mbit/s. In some organizations
the thresholds can be lower than that used in this example; in the event of link
failures, there needs to be spare capacity for traffic to be re-routed.
Some WAN links can be provisioned on top of layer 2 services such as Frame Relay
and ATM; the router-to-router link is a virtual circuit, which is subject not only to
a physical capacity, but also a “logical capacity” limit. The installer or administrator
needs to get the physical link capacity and the QoS parameters. The important QoS
parameters are CIR (committed information rate) for Frame Relay, and MCR
(maximum cell rate) for ATM.
The difference between the current capacity and its acceptable limit is the available
capacity. For example a T1 link used at 48% during the peak hour, with a planning
limit of 85% has an available capacity of about 568 kbit/s.
Comentarios a estos manuales