Agnar Krarup Erlang was born on 1 January 1878 at Lomberg in Jutland. Erlang was introduced to a study of the application of probabilities to telephone problem by Dr Johanssen, who had written two essays on the subject in 1907 and 1908. Erlang was engaged by Dr Johannssen to work with the Copenhagen Telephone Company in 1908, as a scientific collaborator and it was this position which gave him the opportunity to develop and utilise his gifts for the rest of his life. Erlang's most important work "Solution of some problems in the theory of probabilities of significance in automatic telephone exchanges" was published in 1917. This paper contained his formulae for congestion loss and waiting time which are of fundamental importance to the theory of telephohe traffic. Erland died on 3rd February 1929, aged only 51.
Prior to 1943 the name of the unit of telephone traffic had many different variations and in 1943 it was proposed that the name of "Erlang" be introduced as the title of the unit of traffic. This proposal was adopted in Scandinavian countries in 1944 and in 1946 it was proposed to the CCIF (Le Comite Consultatif International des Telephoniques a Grande Distance) and unanimously carried. The following definition is recorded:
The number of calls in progress varies from moment to moment, the actual number depending on the times of arrival and the duration of calls. The above definition implies that the average number of calls in progress simultaneously is numerically equal to the average traffic intensity in erlangs.
The product of the number of calls originating in a given period and the average holding time in this period, expressed in hours, is equal to the traffic volume in erlang hours. If a switch is engaged onone call for 60 minutes its occupany is one, the traffic intensity is one erlang and it carries a traffic volume of one erlang-hour. Similary, twenty 3-minute calls occupying a switch in succession would correspond to the same traffic intensity and volume.
Erlang's theories and later developments by other notable people, have brought systematic rational processes into the "dimensioning" of our networks. "Dimensioning" is the procedure for deciding upon the configuration and number of switches, circuits and other devices needed to handle the traffic intensity with a satisfactory grade of service in the most economical way.
From the Telecommunication Journal of Australia, Vol 25, No 1, 1975.