Internal Addressing for AAL2 user traffic for transport across an ethernet switch
Original Publication Date: 2004-Sep-25
Included in the Prior Art Database: 2004-Sep-25
Related PeopleOther Related People:
The problem to be solved is how to transport user plane protocol frames that arrive via AAL2 (ATM adaptation layer 2) on an ATM (Asynchronous Transfer Mode) line card across a GE (Gigabit Ethernet) ethernet switch to other cards of the network element. Up to now, the problem of transporting UP frames is usually solved by taking the user plane frame, put a proprietary internal address (several bytes long) in front of the header and then put the MAC (Media Access Control) address of the destination card (for ethernet) in front of it. The internal address would identify this direction of the connection uniquely. The new idea to solve the problem is to use the address stack as shown on the right side of figure 1 for transporting UP frames. In front of the user plane frame, on first puts a standard RTP (Real Time Protocol) header and then a standard IP (Internet Protocol) header. Finally, the MAC address of the other card is added. This packet reaches the other card via an ethernet switch. The ATM/AAL2 connection data (VPC (Virtual Path Connection) / VCC (Virtual Channel Connection) / CID (Connection Identification)) identify the connection uniquely at the ATM/AAL2 layer) would be mapped to the UDP (User Datagram Protocol) port number or alternatively to the part of IP address and UDP port number. In the other direction, the MAC header and then the RTP/UDP/IP stack are removed and the UP frames are then put into the ATM/AAL2 format. The UDP port number (or alternatively the pair of IP address and UDP port number) is mapped to the ATM/AAL2 connection data (VPC/VCC/CID). This assumes that each direction of the connection gets assigned a system wide unique UDP port or alternatively a system wide pair of IP address and UDP port number.