High Availability VLAN Overview Configuring High Availability VLANs
page 5-4 OmniSwitch AOS Release 7 Network Configuration Guide June 2013
High Availability VLAN Overview
High availability (HA) VLANs send traffic intended for a single destination MAC address to multiple
switch ports. An HA VLAN is configured by creating a standard VLAN and then assigning ports to the
VLAN. Once these types of ports are assigned, the standard VLAN automatically becomes an HA VLAN.
When this occurs, standard VLAN commands no longer apply.
Destination MAC addresses (unicast and multicast) are also assigned to high availability VLANs. These
addresses identify ingress port traffic that the switch will send out on all egress ports that belong to the
same VLAN
In addition to assigning ingress and egress ports, tagging inter-switch link ports with an HA VLAN ID is
allowed. Ingress port traffic destined for an HA VLAN MAC address is sent out on all egress and inter-
switch link ports that belong to the same VLAN. Traffic forwarded on inter-switch link ports is done so in
accordance with the Spanning Tree state of the port.
A high availability VLAN hosts multiple instances of applications like e-commerce applications, critical
databases, business applications etc and supports redundancy. Each instance may get all the service
requests and based on a shared algorithm, HA VLAN decides on which requests a particular node has to
handle. Apart from service request paths, the nodes are internally connected to share information related to
the service load information, service request data and service availability on other nodes.
The HA VLAN feature on the OmniSwitch provides an elegant and flexible way to connect the server
cluster nodes directly to the ingress network. This involves multicasting the service requests on the config-
ured ports. The multicast criteria is configurable based on destination MAC and destination IP address.
Egress ports can be statically configured on a server cluster or they can be registered by IGMP reports.
The server cluster feature on the OmniSwitch multicast the incoming packets based on the server cluster
configuration on the ports associated with the server cluster.
High Availability VLAN Operational Mode
There are typically two modes of implementation of server clusters in HA VLAN.
• Layer 2 - The server cluster is attached to a L2 switch on which the frames destined to the cluster MAC
address are to be flooded on all interfaces. For more information see “Example 1: Layer 2 Server Clus-
ter” on page 5-9
• Layer 3 - The server cluster is attached to a L3 switch on which the frames destined to the server clus-
ter IP address are to be routed to the server cluster IP and then flooded on all interfaces. For more infor-
mation see “Example 2: Layer 3 Server Cluster” on page 5-11.
Note. The L2 mode is currently supported in AOS using the static mac-address command and L3 mode by
the static ARP command.
To Next Page
Loading...
Need help?
General