|
||||||||||||||
|
|
|||||||||||||
At times our telephone lines become congested and, say, a call to Adelaide may go through Sydney and Melbourne. Sounds sensible, but think for a moment what happens if a large number of calls use an alternate route as opposed to a direct one. The system becomes unnecessarily overloaded. Calls may be using three or more lines instead of one. Worse still new calls that come into an already overloaded system will probably be rerouted and the problem escalates. A "high blocked state" develops and the likelihood of a call being accepted is low. Avoiding the onset of a "high blocked state" is obviously desirable. Traffic congestion can be modelled mathematically. Both Telstra and Optus employ a host of mathematicians in their research and development sections and have close links with research groups within Australian Universities and CSIRO who are working on communications design. At the University of Queensland I have been modelling the effect alternative routing has on network performance. Alternative routing induces a phenomenon called "bistability", that is, the system fluctuates unpredictably between low and high blocking states. Our research has shown that this problem can be tackled using a strategy called "trunk reservation". By reserving small amounts
of capacity for calls using direct routes, bistability can be significantly
reduced and sometimes even removed completely. The result is a predictable
system which can be maintained, for the most part, in a low blocked state. |
||||||||||||||
![]() |
||||||||||||||
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |