Hop-by-hop Performance Measurement Scenarios

In end-to-end performance measurement, only the traffic entering and leaving the network is measured. This measurement reflects only the quality of the entire network. If a network fault occurs, end-to-end performance measurement cannot help locate the fault. To locate the fault, IP FPM provides hop-by-hop performance measurement. The statistics objects in hop-by-hop performance measurement scenarios are the same as that in end-to-end performance measurement scenarios.

On the mobile backbone IP RAN shown in Figure 1, multiple NEs are deployed, and services are complex. Once a network fault occurs, it is difficult to locate the fault.

Figure 1 IP FPM fault locating networking

IP FPM can function with the NMS for fault locating. The process is as follows:

The NMS provides visualized service paths for target flows by segmenting the service forwarding path into multiple closed hops and delivering these closed hops to the DCPs and MCP.
The DCPs report the hop-by-hop measurement data to the MCP.
The MCP calculates the packet loss and delay performance for each hop.
The NMS displays the real-time data of each hop using the MIB, helping locate the fault.

Figure 1 lists how to deploy IP FPM on the network shown in Figure 1.

**Table 1** IP FPM deployment in a fault locating scenario
Deployment Object	IP FPM Deployment
TLP deployment	Hop-by-hop performance measurement applies only to unidirectional target flows. Deploy TLPs as shown in Figure 1 for upstream traffic. Hop-by-hop performance measurement differs from end-to-end performance measurement in the following aspects: TLPs must be deployed hop by hop, but not only on the network ingress and egress. Atomic Closed Hops (ACHs) must be configured, and TLP in-groups and out-groups must be specified for target flows in the ACHs. As shown in Figure 1, the ingress TLPs (TLP0-1 and TLP0-2) and the egress TLPs (TLP1-1 and TLP1-2) can form an ACH. The MCP calculates the packet loss and delay of each closed hop based on hop and synchronization information to obtain hop-by-hop performance of the entire network. Similarly, TLP1-1 and TLP1-2 form an ACH with TLP2-1 and TLP2-2; TLP2-1 and TLP2-2 form an ACH with TLP3-1 and TLP3-2.
DCP deployment	Configure all devices that have TLPs deployed as DCPs to send measurement data to the MCP.
MCP deployment	If routes are reachable between the access and aggregation networks, deploy the MCP on an RSG. If routes are unreachable between the access and aggregation networks, deploy the MCP on an AGG. On the network in Figure 1, deploy the MCP on RSG1.
Clock deployment	Configure the network time protocol (NTP) or 1588v2 so that all device clocks can be synchronized.

ACH Division Based on Measurement Sections

The MCP calculates hop-by-hop performance statistics based on hops along a service path. Accurate measurement can be obtained based only on closed hops. In an ACH, traffic from the ingress TLPs is sent to the egress TLPs, and the egress TLPs receive only traffic sent from the ingress TLPs. As shown in Figure 2, the TLPs on nodes A and B form an ACH with the TLP on node D; the TLPs on nodes D and E form an ACH with the TLP on node F; the TLP on node F forms an ACH with the TLPs on nodes G, H, and I.

Figure 2 Path-based ACH division

In some situations, the accurate path diagram cannot be obtained, and subsequently ACHs cannot be formed. In this case, some key points on the path can be pinpointed to form measurement sections through which traffic passes through, as shown in Figure 3. ACHs can be divided based on these measurement sections.

Figure 3 Measurement section-based ACH division

In hop-by-hop performance measurement, the MCP measures the packet loss and delay based on ACHs. The smaller scale an ACH covers, the more accuracy for fault locating. ACH division helps identify a local area, a direct link, or inbound and outbound interfaces on a device.