The system and method allow Customer Premises Equipment Network Access Devices using ATM as a datalink layer transport, to automatically determine what VC addresses are supported by the remote Network Access Concentrator end of the physical connection. A "divide and conquer" approach is used to ping as quickly as possible the most likely VPI/VCI values. The VC discovery mechanism attempts to discover active VP's and pings only VC's on these VP's. If there are no VP replies, it pings VC's to a limited VCI depth covering the entire VP range. The maximum predetermined range of VC's to ping (MAX-PING) and the physical upstream line speed determines the overall amount of time required by this procedure. Direct benefits of the proposed approach include: time and money saved by not having to customize the firmware in each NAD destined for particular Network Service Providers, and improved end-user experience by providing plug-and-play convenience out of the box.