Specifications

ALAXALA Ring Configuration Guide (Edition 1)
Copyright © 2008, ALAXALA Networks Corporation. All rights reserved.
28
(5) The mode and parameters must be considered when both Ring Protocol and link aggregation are used.
When you use the link aggregation on the port specified as a ring port, note the following:
(a) We recommend that you use the static link aggregation mode.
The link aggregation has two modes, LACP and static, but we recommend that you use static mode, based on
the following reasons:
- Fallback time superiority when a line failure occurs
(In general, compared to LACP mode, which detects unreachable LACPDUs, static mode provides shorter
communication down time when a line failure occurs or is resolved because static mode directly monitors the
link-down state to detect a failure. This difference also affects the switching time by the Ring Protocol because of the
reason following in (b).)
- Direct connection between devices, which is generally used in a local area network
(If devices are directly connected, you can detect a line failure by simply checking for down links on the
devices.)
These are the reasons we recommend you to use static mode.
Also, we recommend that you set link debounce for an Ethernet port to the smallest possible value as long as
there is no particular problem (with line quality, for example).
(b) The parameters for the ring must be set to longer values than the fallback/switching time of the link
aggregation.
If the failure monitoring time of the master node for the Ring Protocol is shorter than the time required for
the fallback or switching operation to complete through the link aggregation, the master node mistakenly
detects a ring failure and switches the path in the ring. As a result, a loop might occur.
More specifically, for the health-check holdtime in Ring Protocol setting, at least the following two criteria
must be satisfied:
i) health-check holdtime > health-check interval
ii) health-check holdtime > link debounce time + 1200 (for the SFP port)
health-check holdtime > link debounce time + 1500 (for the UTP port)
For details on the configuration, see 5.2 Parameter recommendations for chassis- and box
-type switch mixed
networks.
(6) Path (load) balancing through VLAN groups can be achieved.
This is a simple configuration that performs only Layer 2 forwarding, but the path in the ring can be
distributed (load balanced) among VLANs.
You can configure load balancing in the following two ways, which have different features:
(a) Combining two single rings
- The master node can be load balanced.
- Two ring IDs are used (that is, two control VLANs are required and control frames for two rings are sent).
(b) One blocking point for each group in a single ring
- Only one ring ID is used.
- This method has only one master node.
We recommend that you use method (b) if you do not have a particular need for load balancing of the master
node, though the method you select depends on your other configuration requirements.
The following table shows the configuration differences when a single master node is used.
(a) Using two single rings (b) Using one single ring
Configuring the ring port #1 (FL1)
interface gigabitethernet 0/1
switchport mode trunk
switchport trunk allowed vlan 2-3,101-105,106-110
axrp-ring-port 1
axrp-ring-port 2
axrp-primary-port 1 vlan-group 1
interface gigabitethernet 0/1
switchport mode trunk
switchport trunk allowed vlan 2,100-105,106-110
axrp-ring-port 1
axrp-primary-port 1 vlan-group 1