DELL PC7024 User Manual

Port Channel Commands

The hashing algorithm is configurable for each LAG. Typically, an 

administrator is able to choose from hash algorithms utilizing the following 

attributes of a packet to determine the outgoing port: 

• Source MAC, VLAN, EtherType, and incoming port associated with the 

packet. 

• Source IP and Source TCP/UDP fields of the packet. 
• Destination MAC, VLAN, EtherType, and incoming port associated with 

the packet. 

• Source MAC, Destination MAC, VLAN, EtherType, and incoming port 

associated with the packet. 

• Destination IP and Destination TCP/UDP Port fields of the packet. 
• Source/Destination MAC, VLAN, EtherType, and incoming port 

associated with the packet. 

• Source/Destination IP and source/destination TCP/UDP Port fields of the 

packet.

PowerConnect devices based on Broadcom XGS-IV silicon support 

configuration of hashing algorithms for each LAG interface.  The hashing 

algorithm is used to distribute traffic load among the physical ports of the 

LAG while preserving the per-flow packet order.
One limitation with earlier LAG hashing techniques is that the packet 

attributes were fixed for all type of packets.  Also, there was no MODULO-N 

operation involved, which can result in poor load balancing performance.
As part of Release 4.0, the LAG hashing support is extended to support an 

Enhanced hashing mode, which has the following advantages:

• MODULO-N operation based on the number of ports in the LAG.
• Packet attributes selection based on the packet type.  For L2 packets, 

Source and Destination MAC address are used for hash computation.  For 

IP packets, Source IP, Destination IP address, TCP/UDP ports are used.

• Non-Unicast traffic and Unicast traffic is hashed using a common hash 

algorithm.

• Excellent load balancing performance.

2CSPC4.XCT-SWUM2XX1.book  Page 626  Monday, October 3, 2011  11:05 AM