CPE ping

The Nokia-specific CPE ping function provides a common approach to determine if a destination IPv4 address can be resolved to a MAC address beyond the Layer 2 PE, in the direction of the CPE. The function is supported for both VPLS and Epipe services and on a number of different connection types. The service type determines the packet format for network connection transmissions. The transmission of the packet from a PE egressing an access connection is a standard ARP packet. This allows for next-hop resolution for even unmanaged service elements. In many cases, responses to ICMP echo requests are restricted to trusted network segments only; however, ARP packets are typically processed.

If the ARP response is processed on a local SAP connection on the same node from which the command was executed, the detailed SAP information is returned as part of the display function. If the response is not local, the format of the display depends on the service type.

The VPLS service construct is multipoint by nature, and simply returning a positive response to a reachability request would not supply enough information. For this reason, VPLS service CPE ping requests use the Nokia-specific MAC ping packet format. Execution of the CPE ping command generates a MAC ping packet using a broadcast Layer 2 address on all non-access ports. This packet allows for more information about the location of the target. A positive result displays the IP address of the Layer 2 PE and SAP information for the target location.

Each PE, including the local PE, that receives a MAC ping proxies an ARP request on behalf of the original source, as part of the CPE ping function. If a response is received for the ARP request, the Layer 2 PE processes the request, translates the ARP response, and responds back to the initial source with the appropriate MAC ping response and fields.

The MAC ping OAM tool makes it possible to detect whether a particular IPv4 address and MAC address have been learned in a VPLS, and on which SAP the target was found.

The Epipe service construction is that of cross-connection, and returning a positive response to a reachability request is an acceptable approach. For this reason, Epipe service CPE ping requests use standard ARP requests and proxy ARP processing. A positive result displays remote-SAP for any non-local responses. Because Epipe services are point-to-point, the path toward the remote SAP for the service should already be understood.

Nokia recommends that a source IP address of all zeros (0.0.0.0) is used, which prevents the exposure of the provider IP address to the CPE.

The CPE ping function requires symmetrical data paths for correct functionality. Issues may arise when the request egresses a PE and the response arrives on a related but different PE. When dealing with asymmetrical paths, the return-control option may be used to bypass some of the asymmetrical path issues. Asymmetrical paths can be common in all active multi-homing solutions.

For all applications except basic VPLS services (SAP and SDP bindings without a PBB context), CPE ping functionality requires minimum FP2-based hardware for all connections that may be involved in the transmission or processing of the proxy function.

This approach should only be considered for unmanaged solutions where standard Ethernet CFM (ETH-CFM) functions cannot be deployed. ETH-CFM has a robust set of fault and performance functions that are purpose-built for Ethernet services and transport.

Connection types used to support VPLS and Epipes include SAPs, SDP bindings, B-VPLS, BGP-AD, BGP-VPWS, BGP-VPLS, and MPLS-EVPN.