User Manual

34
III-3-2. REDUNDANCY – ERPS
Ethernet Ring Protection Switching (ERPS) applies the protection switching mechanism for Ethernet
traffic in a ring topology. This mechanism is defined in ITU-T G8032. You can avoid the possible loops in
a network by implementing the ERPS function. This is done by blocking the flow of traffic to the Ring
Protection Link (RPL) for protecting the entire Ethernet ring.
When an ERPS is implemented in a ring topology, only one switch is allocated as the owner. This
switch is in charge of blocking the traffic in the RPL to avoid loops. The switch adjacent to the RPL owner
is called the RPL neighbor node and it is responsible for blocking the end of the RPL during normal
condition. The participating switches that are adjacent to the RPL owner or neighbor in a ring are called
the members or RPL next-neighbor nodes. The primary function of these switches is to forward the
received traffic.
To make sure that a ring is up and loop-free, Ring Automatic Protection Switching message is sent
regularly as control messages by nodes on the ring. The RPL owner identifies a signal failure (SF) in a
ring when the RPL owner misses the poll packets or reads from the fault detection packets. When the
fault is identified, the RPL owner unblocks the ring protection link (RPL) and permits the protected VLAN
traffic through.
ERPS, similar to STP, provides a loop-free network by using polling packets to detect faults. If a fault
occurs, ERPS restores itself by sending traffic over a protected reverse path rather than making a
calculation to identify the forwarding path. The fault detection mechanism in the ERPS enables the ERPS
to join in less than 50 milliseconds and recovers quickly to forward traffic.