CANOGA PERKINS 9145E NID Software Version 4.10 - Manual (Page 100)

9145E NID Software User’s Manual Diagnostics
Network Performance
88
PLPM partner NIDs are aware of their partner Manager IP address and MAC address. Any
management traffic received by one NID that is destined to its partner will be forwarded to the
partner through its MPP.
4.6.15.2 Determining the Port to Forward Test Traffic
It is important that synthetic test traffic follow the same path through the network as the EVCs
being tested. This holds true for the LAG protected interface shown in Figure 4-42. There are
two types of LAG that a Service Provider might use for protection:
Say that the Active Master NID has a PM test instance scheduled for each of the EVCs traversing
the interface in Figure 4-42. The Active Master NID needs to determine which of the two links the
EVC traverses in the LAG in order to determine where to send the synthetic Test Packet it
generates. If a Test Packet is generated for an EVC that traverses the link that the Active Master
NID is in-line with, then it forwards the packet out its Network Port. If a Test Packet is generated
for an EVC that traverses the link that the Passive Master NID is in-line with, then it forwards the
packet out its MPP to the Passive Master, so the Passive Master can in-turn forward it out its
Network Port. In order for the Active Master NID to determine the Test Packet forwarding
scheme, it must first determine whether there is an Active/Passive or Active/Active LAG.
The NIDs “snoop” the LACP exchanged between the two switches in Figure 1 in order to
determine whether there is an Active/Passive LAG or Active/Active LAG. The NIDs do not peer
the LACP protocol, but simply intercept and interpret the packets exchanged between the two
switches to determine whether the link they are on is currently in the “active” or “standby” state.
Once each NID identifies the state of the link they are on, they exchange that state information to
their PLPM Partner NID through PLCP. The Active Master NID then determines the Test Packet
forwarding scheme based on this information:
• If the Active Master NID's link is forwarding frames and the Passive Master NID's link is
not forwarding frames, the LAG is in Active/Passive mode. In this case, the Active Master
NID will forward all Test Packets generated out its Network Port.
• If the Active Master NID's link is not forwarding frames and the Passive Master NID's link
is forwarding frames, the LAG is in Active/Passive mode. In this case, the Active Master
NID will forward all Test Packets generated out its MPP towards the Passive Master NID,
so that the Passive Master may, in turn, forward these out its Network Port.
• If the Active Master NID's link and the Passive Master NID's link are both forwarding
frames, the LAG is in Active/Active mode. In this case, the Active Master NID will refer-
Active/Passive LAG All the traffic is forwarded by the LAG function to one of two ports in the
two-port LAG (i.e., traffic is forwarded to the “active” port, while the
“passive” port remains idle). If this port experiences a fault, all the
traffic will be “protection switched” to the other port in the LAG.
Therefore, all the traffic is being forwarded on one of the two ports at
any given time.
Active/Active LAG Each port in a two-port LAG forwards about half of the LAG traffic.
Which EVCs are forwarded by which port in the LAG is determined by
the LAG frame distribution function, which is a proprietary function
implemented by switches. If there is a fault on one of the links, the
other link will begin forwarding all frames, effectively doubling the traffic
on the link.