Lucent Technologies Octel Unified Messenger, Planning Manual

Page: 30 / 120

Octel Unified Messenger Concepts & Planning Guide
2-4 101-1620-005
Sizing ports
Accurate port sizing is critical to the successful implementation of any
voice processing application. A careful analysis of system traffic patterns
must be made. Specifically, the volume of calls generated by internal
subscribers and external callers to the system during the busiest hour of
the day must be determined.
Busy hour traffic is used in the calculation to determine the number of
ports needed to provide an acceptable level of service to the users. Most
newer PBXs provide traffic statistic reports that can provide an accurate
picture of traffic patterns on the switch. When preparing the
implementation of a messaging system, the planner should study a
minimum of one week’s traffic data to determine daily and hourly call
volumes.
How to determine
the busy hour
Most PBXs can generate traffic reports that provide statistics on a weekly,
daily, or hourly basis. Usually these reports break out the traffic statistics
by type of call, for example, incoming calls, outgoing calls, and calls to
specific hunt groups. These reports can be used to determine specific
traffic patterns and when the busy hour occurs.
Suppose that during the course of a typical business week the traffic
report indicates that Monday is the busiest day for incoming calls. On
Monday a total of 3,500 calls are received, and the hourly statistics show
that the busiest hour of the day occurs between 9 a.m. and 10 a.m., with
350 incoming calls. Therefore 9 a.m. to 10 a.m. is the customer’s busy
hour.
Another way of expressing busy hour traffic is as a percentage of total
daily traffic. In the example above, 350 busy hour calls is 10% of the
3,500 total calls received for the day. If traffic statistics are not available,
then an educated guess at busy hour traffic is required. The tables in
Appendix A, “Sizing for Ports” can be used as guidelines for estimating
how many users can be supported with a given number of ports based on
some assumptions about average daily port usage per subscriber. These
tables also provide for traffic patterns of 10%, 14%, or 18% busy hour
traffic.