| 摘要 |
A method and apparatus is disclosed wherein a cost of bandwidth usage in a virtual private network is calculated as a function of a plurality of traffic matrices associated with said bandwidth. In a first embodiment, the cost is calculated as a function of the maximum number of channels required to support an upper bound of the bandwidth of all connections originating from at least one node in the network. In accordance with another embodiment, the cost of providing service is calculated as a function of a traffic matrix in said plurality of traffic matrices, said traffic matrix requiring the highest possible bandwidth use of all traffic matrices in said plurality of traffic matrices. Finally, in accordance with a third embodiment, said cost is calculated as a function of the most-likely amount of bandwidth used by a customer.
|
