Web Management Guide-R06

Chapter 13 | Basic Administration Protocols
Connectivity Fault Management
– 496 –
Figure 306: Multiple CFM Maintenance Domains
Note that the Service Instances within each domain shown above are based on a
unique maintenance association for the specific users, distinguished by the domain
name, maintenance level, maintenance association’s name, and assigned VLAN.
Basic CFM Operations
CFM uses standard Ethernet frames for sending protocol messages. Both the source
and destination address for these messages are based on unicast or multicast MAC
addresses, and therefore confined to a single Layer 2 CFM service VLAN. For this
reason, the transmission, forwarding, and processing of CFM frames is performed
by bridges, not routers. Bridges that do not recognize CFM messages forward them
as normal data. There are three basic types of CFM messages, including continuity
check, link trace, and loop back.
Continuity check messages (CCMs) are multicast within a single Service Instance
(i.e., a specific MA), allowing MEPs to discover other MEPs within the same MA, and
MIPs to discover MEPs. Connectivity faults are indicated when a known MEP stops
sending CCMs, or a remote MEP configured in a static list does not come up.
Configuration errors, such as a cross-connect between different MAs, are indicated
when a CCM is received with an incorrect MA identifier or maintenance level.
Loopback messages are used for fault verification. These messages can be sent
using the MAC address of any destination MEP within the same MA. If the target
MEP’s identifier has been discovered through CCM messages, then a loop back
message can also be sent using the MEP’s identifier. A reply indicates that the
destination is reachable.
Link trace messages are used for fault verification. These messages are multicast
frames sent out to track the hop-by-hop path to a target MEP within the same MA.
Responses provide information on the ingress, egress, and relay action taken at
each hop along the path, providing vital information about connectivity problems.
Responses allow the sender to discover all of the maintenance points that would be
traversed by a data frame sent to the target MAC address.
SNMP traps can also be configured to provide an automated method of fault
notification. If the fault notification generator detects one or more defects within
Customer MA
Provider MA
Operator 1 MA Operator 2 MA
C
C
C
C
O
1
O
1
P
O
2
O
2
P
P
O
1
O
2
P