Allied Telesis x900-48FE-N User Manual

Page 5 | AlliedWare™ OS How To Note: EPSR

EPSR uses a fault detection scheme that alerts the ring 
when a break occurs, instead of using a spanning tree-
like calculation to determine the best path. The ring 
then automatically heals itself by sending traffic over a 
protected reverse path.

EPSR uses the following two methods to detect when 
a transit node or a link goes down:

•

Master node polling fault detection

To check the condition of the ring, the master 
node regularly sends Health messages out its 
primary port, as described in 

. If all links and nodes in the ring are 

up, the messages arrive back at the master node on 
its secondary port. 

This can be a relatively slow detection method, 
because it depends on how often the node sends 
Health messages.

Note that the master node only ever sends Health 
messages out its primary port. If its primary port 
goes down, it does not send Health messages.

•

Transit node unsolicited fault detection

To speed up fault detection, EPSR transit nodes 
directly communicate when one of their interfaces 
goes down. When a transit node detects a fault at 
one of its interfaces, it immediately sends a Link- 
Down message over the link that remains up. This 
notifies the master node that the ring is broken and 
causes it to respond immediately.

When the master node detects an outage somewhere 
in the ring, using either detection method, it restores 
traffic flow by:

1.

declaring the ring to be in a Failed state

2.

unblocking its secondary port, which enables data 
VLAN traffic to pass between its primary and 
secondary ports

3.

flushing its own forwarding database (FDB) for the 
two ring ports

4.

sending an EPSR Ring-Down-Flush-FDB control message to all the transit nodes, via 
both its primary and secondary ports

The transit nodes respond to the Ring-Down-Flush-FDB message by flushing their 
forwarding databases for each of their ring ports. As the data starts to flow in the ring’s 

The state when there are no link or 
node failures on the ring.

The state when there is a link or 
node failure on the ring. This state 
indicates that the master node 
received a Link-Down message or 
that the failover timer expired before 
the master node’s secondary port 
received a Health message.

The state when EPSR is first 
configured, before the master node 
determines that all links in the ring 
are up. In this state, both ports on 
the node are blocked for the data 
VLAN. From this state, the node can 
move to Links Up or Links Down.

The state when both the node’s ring 
ports are up and forwarding. From 
this state, the node can move to 
Links Down.

The state when one or both of the 
node’s ring ports are down. From this 
state, the node can move to Pre-
forwarding

The state when both ring ports are 
up, but one has only just come up and 
is still blocked to prevent loops. From 
this state, the transit node can move 
to Links Up if the master node blocks 
its secondary port, or to Links Down 
if another port goes down.