Hi byrdy96.
The Source/Destination MAC addresses the ACSA guide is referring to - when the LAG's hashing setting is configured to use an L2 hashing mechanism, as example (or the Source/Destination IP Addresses, when the LAG's hashing setting is configured to use an L3 hashing mechanism capable of failing back to the L2 hashing mechanism just in case no L3 Src/Dst data from peer is available) - refer to the Source/Destination MAC Addresses of peers involved in the communication (direction: Source to Destination), they don't refer to the LAG Member Interfaces' MAC Addresses.
So the LAG's member interface selected by the L2/L3 Hashing algorithm (mechanism) for traffic leaving the LAG logical interface is selected by using the Source/Destination L2 data or Source/Destination L3 data (fallback to L2 when necessary) provided by the peers involved in each communication, the less are the variations the hashing algorithm can use to select the egressing physical interface (Source MAC/IP don't change? Destination MAC/IP don't change? both Source and Destination MAC/IP don't change?) the higher is the probability the traffic is biased (polarization) to use a particular LAG's member interface over the other(s).
Original Message:
Sent: Jul 26, 2022 04:21 AM
From: Kurtis Byrd
Subject: Aruba CX LAG Interfaces
I am currently working through the content for the ACSA Certification. Looking at the LAG Interface section within the ACSA Study Guide, i am getting a bit confused with the Load Sharing/Balancing feature of LAG. The book mentions the switch uses a Hash algorithm to determine whicch port is used for a data packet within a LAG. It either uses the Source/Destination MAC Address combination as an input for the Hash or Source/Destination IP Address to determine the port used as the output, however all ports in a LAG share the same Source addresses which has me confused - what is different to port 1 vs port 2 in a LAG?