An understanding of the basic principals of network design, will allow a telecommunications manager to optimize available resources. Equipped with an understanding of queuing theory and the ability to use traffic tables are an asset in gathering data to answer many telecommunication questions. Calculating traffic formulas, with the use of traffic tables or computer applications (such as like the ACD Analyzer), the manager can reach a balance between costs and service.
Basic Principals that Govern Network DesignListed below are the four basic principals
that govern network design:
1.
Large groups of servers are more efficient
than smaller groups - this is commonly called occupancy
2.
For a given amount of traffic load,
the greater the number of servers, the less likely it is that callers will have
to wait
3.
The rate at which calls arrive is
seldom uniform - there is a wide variation in the number of calls
4.
Alternate routes to the destination
increase the robustness of a network - traffic can overflow to an alternate
route
In order to administer a telecommunications system, the following are needed:
Grade
of service – the percentage of calls that encounter some form
of blockage
·
If too high
– many circuits will carry little or no traffic
·
If too low
– many circuits will be busy and productivity will decrease
·
Most circuits
are designed for grade of service between 1 and 5 percent blockage
·
A way of increasing
the load on a group of circuits is through alternate routing
Traffic
load – expressed as the amount of traffic presented to
a trunk group in its busiest hour
·
Measured in
minutes, hours, hundreds of call seconds (CCS) or erlangs.
·
Expressed as
the relationship between the calling rate and length of an average call
Busy hour – the average amount of traffic that flows during the 10 highest days of the year
Offered
verses carried load -
·
The difference
between the two lies in the time required for call setup and handling incomplete
and blocked calls
·
Long distance
call statements must be adjusted to convert billed minutes into usage minutes
Quantity of trunks – can be calculated from traffic tables when the load and grade of service are known
Using
Erlang C tables – used when callers wait in queue
·
Limitations
of Erlang C – it does not provide information about the service received by
callers are not within the objective service level range and it assumes that
calls arrive randomly
·
Utilization
– the percentage of time agents are activity handling calls either in talking
or in wrap-up mode
Calculating voice mail port requirements – calculate using Erlang B tables
Using
an ACD Analyzer for traffic calculations
– this computer application simplifies the process
·
Two modes
·
Trunking using
the Erlang B formula
·
Queuing using
the Erlang C formula
Busy signals may send the wrong message
to your callers. The alternative is typically waiting in a queue.
·
Busy signals
can be returned in two ways
·
Reduce the
number of trunks, allowing the carrier to provide the busy signal
·
Provide the
busy signal from the ACD
·
Busy signals
can tell you if a caller returns if the call center gathers automatic number
identification with each call
The following are sources of traffic
usage data and methods used to interpret and adjust traffic usage:
·
Switching system
data – readings can be obtained in raw form or a GUI-based program
·
Usage-sensitive
bills – local and toll billing statements
·
LEC traffic
or busy study – LECs can perform a busy study
·
Call accounting
system – reasonably accurate usage information
Case Study: Using Traffic Theory to Optimize Local Trunks
In a case study involving the review
of bills from a LEC and an IXC, it is evident that the PBX has trunking problems.
·
There are not
enough two-way central office trunks
·
Several assumptions
are made in order to correct the problems
·
When rearranging
trunks, telecommunications managers must review assumptions carefully
In summary, the use of traffic tables and computer applications are an asset to the telecommunications manager. He/she must make many assumptions and the data is often inaccurate. By frequently monitoring the data and making trunk adjustments, the data will be more accurate.