Brocade Network OS Administrator’s Guide v4.1.1 - Manual (Page 147)

Brocade trunks
Network OS 4.0.0 and later supports Brocade trunks (hardware-based link aggregation groups, or
LAGs). These LAGs are dynamically formed between two adjacent switches. The trunk formation is
controlled by the same Fibre Channel Trunking protocol that controls the trunk formation on FC
switches. As such, it does not require user intervention or configuration except enabling or disabling,
which instructs the switch software to form a trunk at the global level or not. All ISL ports connected to
the same neighbor Brocade switch will attempt to form a trunk. Refer to
Enabling a fabric trunk on page
151 for instructions.
Ports groups have been established on supported standalone switches and on line cards in chassis
systems for trunking. For a successful trunk formation, all ports on the local switch must be part of the
same port group and must be configured at the same speed. Following are the number of ports allowed
per trunk from port groups in supported platforms. For details on how port groups are arranged on these
platforms, refer to the switch or chassis system Hardware Reference Manual.
• VDX 6720 and 6730 - up to eight ports allowed per trunk. - eight ports per trunk.
• VDX 6740 switches - up to sixteen ports per trunk.
• 48x10 GbE line card - up to eight ports per trunk.
• 48x10G-T line card - up to 16 ports per trunk.
• 12x40 GbE line card - up to two 40GbE ports are allowed per trunk, and these ports must be
configured in breakout mode.
• 27x40 GbE line card - up to two 40GbE ports are allowed per trunk, and these ports must be
configured in breakout mode. Note that breakout mode is only allowed on the first two ports in port
groups that are configured in Performance operating mode. Refer to the Brocade 8770 Hardware
Reference Manual for more information on line card operating modes.
The following additional rules apply to Brocade trunks:
• On Brocade VDX 6740 switches, low-volume traffic below certain thresholds may not be evenly
distributed on all links. This threshold can be as low as 64K.
• The trunk is turned on by default.
• Trunks are not supported between the Brocade 8000 and the Brocade VDX 8770.
Fabric formation
Brocade VCS Fabric technology leverages proven FC Fabric protocols to build a TRILL fabric. The main
functions of the fabric formation protocols are as follows:
• Assign the Brocade VCS Fabric-wide unique RBridge IDs (Domain ID Assignment).
• Create the network topology database using link state routing protocol (Fabric Shortest Path First, or
FSPF). FSPF calculates the shortest path routes to a destination RBridge.
• Distribute fabric multicast traffic.
Principal switch selection
Every Brocade VCS Fabric-enabled switch, upon boot-up and after the Fabric port formation, declares
itself to be a principal switch and advertises this intent on all fabric ports. The intent includes a priority
and its switch WWN. If all switches boot up at the same time, the default priority is the same and all
switches will compare their mutual intents. The switch with the lowest Switch WWN becomes the
principal switch. The WWN is an industry-standard burnt-in switch identifier, similar to the Bridge-MAC
except it is 8 bytes. The role of the principal switch is to decide whether a new RBridge joining the fabric
conflicts with any of the RBridge IDs already present in the fabric. At the end of the principal switch
selection process, all the switches in the cluster have formed a tree with the principal switch at the root.
Brocade trunks
Network OS Administrator’s Guide 147
53-1003225-04