Brocade Multi-Service IronWare Administration Guide (Supporting R05.6.00) - Manual (Page 251)

Multi-Service IronWare Administration Guide 233
53-1003028-02
Frame Loss Measurement
6
Frame Loss Measurement
The Frame Loss Measurement feature (ETH-LM) maintains counters of received and transmitted
data frames between a pair of MEPs. These counters are used to calculate the frame loss ratio.
Only single-ended ETH-LM, which is used for on-demand OAM, is supported. An MEP sends frames
with an ETH-LM request information to its peer MEP and receives frames with ETH-LM reply
information from its peer MEP to perform loss measurement. Frames which carry the Loss
Measurement Message (LMM) PDU are called LMM frames. Frames which carry the Loss
Measurement Reply (LMR) PDU are called LMR frames.
When the Loss Measurement Message (LMM) is configured the Frame Loss Measurement is
enabled.
LMM over VLAN
Frame Loss Measurement can be done over VLAN where Connectivity Fault Management (CFM) is
configured. In this use case, CFM should be enabled and down MEP should be configured on the
VLAN end-points which should be monitored. LMM can be configured on the end-points for periodic
measurements irrespective of the CFM connectivity. Ensure CFM connectivity is UP and running
before the LMM session actually get started. Otherwise, an error will be thrown.
LMM over VPLS
Frame Loss Measurement can be done over VPLS and VLL where Connectivity Fault Management
(CFM) is configured. In this use case, CFM should be enabled and UP MEP should be configured on
the VPLS end-points which should be monitored. LMM can be configured on the end-points for
periodic measurements irrespective of the CFM connectivity. Ensure CFM connectivity is UP and
running before the LMM session actually get started. Otherwise, an error will be thrown.
Configuration considerations and limitations
As the Frame Loss Measurement feature uses ACL for getting data packet counters, it will be
affected as follows:
• When there is an active LMM session and an L2 ACL is getting bounded, there will be some
drop or frame loss expected as the LMM ACL is getting re-programmed.
• The responder should be started first before starting the initiator. Otherwise, the LMM packets
will be dropped at the responder and no ACLs will be programmed, which may lead to
inaccurate results.
• During termination, stop the initiator before the responder. Stopping the responder first may
lead to inaccurate results as mentioned in the previous point.
• Only one LMM session will be active per source MEP per priority. This means eight active
sessions per source MEP, one active for each priority.
• Maximum of 32 LMM sessions can be created per source MEP (irrespective of the priority).
• Maximum of 100 LMM sessions can be activated per system at any given point of time
irrespective of the MD, MA, and MEP.
• LMM functionality not guaranteed if there exists multiple VPLS end points sharing the single
peer for that VPLS instance. There should be a single VPLS end point.