Table of Contents
Advertisement
Traffic engineering
To determine the anticipated processor occupancy that is associated with a particular
configuration, the average processing cost per call must be determined based on the
anticipated volume of each type of call, and the complexity of the various call types. This
average cost per call implies the slope of the line in relating static and CP occupancy to the
BHCC rate. The intercept of that line, which corresponds to the no-load occupancy, depends on
several factors, including which Communication Manager platform is being used, how many
endpoints are administered, and so on.
Communication Manager systems are designed to keep the sum of static and CP occupancy
below a particular threshold. This is done to allow a suitable amount of processing time for
system management functions.
So for a given configuration, the various types of calls to be supported are identified, and the
processing cost for each call type (based upon the complexity of the call) must be assessed.
That information can then be used to determine the average processing cost per call, based on
the anticipated relative frequencies of the various call types. The slope of the line relating the
sum of static and CP occupancy can then be determined from the average processing cost per
call. The intercept of that line is determined by information such as the Communication
Manager platform used, the number of endpoints administered, and so on.
Therefore, for the given configuration, the specific linear model for the relationship between the
sum of static and CP occupancy, as a function of BHCC, has been derived. Using the
anticipated BHCC rate in that model yields the expected combined static and CP occupancy. If
that value exceeds the preset threshold, the configuration is unacceptable for the anticipated
call rate. In such a case, to support that call rate, either another platform must be considered, or
multiple platforms must be networked together.
IP Bandwidth and Call Admission Control
IP bandwidth analysis for media streams begins with determining the number of bidirectional
media streams that are associated with each type of call supported by the system. Throughout
this discussion, calls between two IP stations are assumed to use shuffling. That being the
case,
Figure 57: Connectivity modes between two IP endpoints
intrasite call between two IP endpoints requires a single bidirectional media stream through the
LAN at that site.
indicates that each intrasite call between an IP endpoint and a circuit-switched endpoint
(including PSTN trunks) also requires a single bidirectional media stream through the LAN at
that site. In addition,
page 164 indicates that each interport network intrasite call between two circuit-switched
endpoints (including PSTN trunks) also requires a single bidirectional media stream through the
LAN at that site (assuming that IP-Connect is used, as opposed to a circuit-switched center
stage). In fact, the only intrasite call that does not require a single bidirectional media stream
through the LAN at that site is an intraport network call between two circuit-switched endpoints
which requires no IP resources because the call is completed solely across the circuit-switched
TDM bus of the PN. Each intersite call requires exactly one bidirectional media stream through
each participating site's LAN, as well as a single bidirectional media stream through the WAN
that connects the two sites.
176 Avaya Application Solutions IP Telephony Deployment Guide
Figure 56: Examples of media streams between Avaya endpoints
Figure 56: Examples of media streams between Avaya endpoints
on page 166 indicates that an
on page 164
on
Advertisement
Table of Contents
