Specifications

ALAXALA Ring Configuration Guide (Edition 1)
Copyright © 2008, ALAXALA Networks Corporation. All rights reserved.
70
5. Setting Various Parameters
5.1 Parameter descriptions
You can specify the following parameters when you configure a ring network.
(1) health-check interval
This parameter specifies the interval for sending health-check frames from master nodes or terminal shared
nodes in a shared link non-monitoring ring.
Chassis-type switch: 10 to 10,000 ms (default: 100 ms)
Box-type switch: 500 to 60,000 ms (default: 1,000 ms)
If you mix two types of switches to build your network and you use a chassis-type switch as the master node,
you need to use the value range for box-type switches that have a longer sending interval of health-check frames.
Using the value range for chassis-type switches might lead to increased CPU load on box-type switches running
as transit nodes. (In a mixed network configuration of chassis- and box-type switches, health-check interval >=
500 ms.)
(2) health-check holdtime
This parameter specifies the hold time until the master node determines that a failure has occurred after the
node receives the last health-check frame sent by the node itself or the terminal shared nodes in the shared link
non-monitoring ring.
Chassis-type switch: 32 to 12,288 ms (default: 256 ms)
Box-type switch: 500 to 300,000 ms (default: 3,000 ms)
If you mix two types of switches to build your network and you use a chassis-type switch as the master node,
you need to use an appropriate value range because the sending interval of the health-check frames has already
been set to the value range for box-type switches. (In a mixed network configuration of chassis- and box-type
switches, health-check holdtime > 500 ms.)
Also, setting this parameter to a smaller value (that is, closer to the value of the health-check interval) allows
you to detect a ring failure earlier. However, setting the parameter to a value less than twice the health-check
interval might result in false detection of a ring failure if a health-check frame is discarded for some reason.
(Recommended value: health-check holdtime >= health-check interval x 3)
(3) forwarding-shift-time
This parameter specifies the reception hold time for flush control frames on transit nodes. After this time
elapses, the transit node clears the MAC address table for the ring port and changes the port status from BLK to
FWD. This action takes place even if the node does not receive flush control frames.
1 to 65,535 sec or infinity (default: 10 sec)
When you specify this parameter, you need to set a larger value than the health-check interval set for the
master node. Otherwise, a loop might occur because the master node performs a failure recovery operation and
changes the port status to FWD at both ends of the failed link before the ring port is changed to BLK status
(forwarding-shift-time >> health-check interval).
(4) flush-request-count
This parameter specifies the number of times the master node sends flush control frames, which direct transit
nodes in a ring to clear their MAC address tables when a ring failure occurs or a failure is recovered from.
1 to 10 (default: 3)
If the flush control frame is discarded on the network before the frame reaches transit nodes, the MAC address
tables on the nodes might not be cleared. Flush control frames must be sent several times.
(5) link debounce
This parameter specifies the link-down detection time until a link-down is detected after the detection of a link
failure.
Chassis-type switch: time 0 to 10,000 ms
(Default: 10/100/1000BASE-T=2,000 ms, 1000BASE-X/10GBASE-R=0 ms)
Box-type switch: time 0 to 10,000 ms (default: 2,000 ms)
When you configure LA on the ring ports, you can change the time until a link fault is detected by tuning the
link debounce value. Therefore, when you configure LA on the ring ports, you need to set an appropriate
health-check holdtime value that corresponds to the link debounce time value and the fallback time of LAs. If the
health-check holdtime value is small, the Ring Protocol might erroneously detect a failure when a line with LA is