Traffic Engineering

When designing, engineering or assessing the network requirements for voice and data circuits, trunks or routes to, from or between an organization’s sites, it is necessary to determine the quantity of facilities based on actual or proposed traffic. 

If the traffic is for existing general business situations, this is accomplished by determining the amount of traffic during the busiest hour of the busiest day of the week.  In voice situations, a ”Busy Studies” can be run and in the data world it would be a utilization study.  Busy Studies can be obtained from the PBX / IPT system or from the local or long distance provider.  Data utilization can be obtained by using various off-the-shelf tools that the organization’s data people most likely would be currently using for monitoring the existing LAN/WAN (e.g., MRTG reports).

When traffic is not available, as would be for proposed or future situations (e.g., during the Christmas holiday season for retail organizations, marketing campaigns such as late summer in the auto industry or during fund-raisers for charitable and political campaigns) the traffic can be determined through forecasts and/or by comparing to past similar situations.

Once the traffic is obtained, it is then necessary to determine which traffic engineering formula (i.e., Erlang table) to use, either Erlang B (standard engineering for normal business situations) or Erlang C (situations involving traffic queuing such as with Call Centers & Outbound calling campaigns).  The Erlang Engineering formulas are similar to Poisson Traffic Engineering, but a little more conservative.  See below for detained explanation of the Erlang formulas.

Using these formulas can be very cumbersome and time consuming if using the Erlang Tables found in books or through the internet or quite easy by using the Traffic Configurator software developed by RMS and found  Not only is using this software quick and simple to use but reports can be generated for presentations and documentation.  Furthermore, the Grade of Service (GoS) for the desired amount of call paths or the quantity of agents necessary in a Queue group based on desired queue length are provided. 

Erlang B

This formula is used to determine the quantity of facilities necessary for a network group based on a desired minimal percentage of blocked calls during the busiest hour (e.g., 1%).  This group of network facilities can consist of Call Paths (such as with VoIP), trunks, lines and circuits between two “End Points”.  These End Points can be PRIs used between a premise-based voice system and a Local Exchange Carrier (LEC) or T-1 voice circuits between two company sites.  Furthermore, this formula assumes that blocked calls (i.e., busies) are lost and not retried.  (There are other less used B-type formulas that assume different situations such as blocked calls are retried again within the Busy Hour.)   While most of us would prefer to never have any blocked calls, in reality, this is unrealistic for two reasons.  First, it would necessitate a significant additional quantity of network facilities that would most likely be cost prohibitive.  Second, telephone companies typically engineer their street facilities based on Erlang B with 1% blockage (i.e., fast-busy tone when making a call); therefore, the street facilities are the funnel point that cannot be bettered.

Erlang C

This formula is typically used in a queuing environment to determine the quantity of Answer Points needed to handle the traffic provided based on a desired maximum length of queue during the Busy Hour.  These answer points typically include system operators, Call Center agents, Help Desk support groups, etc.  Basically, the key aspect of this formula is not based on blocked calls but rather the maximum length of queue / hold-time desired (e.g., 30 seconds) before calls are answered.  This formula also determines the overall amount of traffic in Erlangs or minutes that not only includes the time while on calls but also the hold-time in queue; from which, Erlang B can then be used to determine the quantity of network facilities needed for this Queue group during the Busy Hour.


While the Erlangs are derived from industry-proven engineering formulas, and have an extremely high degree of accuracy, actual real-time usage may vary somewhat, primarily because these formulas are based on statistical averages and/or past call volumes and patterns and do not take into account future deviations and situations (e.g., busy hour of day may differ due to outside occurrences such as disasters, market fluctuations, competitive strengths, etc.).