I did a little miscalculation and forgot to account for the traffic flowing between the peer ports.
I bought two CRS326-24G-2S+ and was planning on using one or two of the SFP±ports as uplinks. However, that’d only leave gigabit ports for the MLAG-interconnect.
If I only use one SFP+ per switch for the interconnect and one for the uplink, assuming actual traffic flows through the interconnect, I’d already have it run at it’s max capacity, leaving nothing for the other 24 gigabit ports.
So my question is: What and how much traffic actually goes through the interconnect - how much did I mess up?
From the tests I carried out assuming an optimal network condition, where all the ports are UP, the traffic passes through the bonding interfaces, without therefore asking the peer port. The traffic on the peer port passes when the client must necessarily do so from the peer port to reach the destination, for example:
The two switches can be seen from the peer port, for example SFP24
If you interrupt the SFP1 / Client1 / Switch 1 and the SFP2 / Client2 / Switch2 you will see that the traffic passes from the SFP24 or the peer port ..
Hello
This is not the behavior I have seen with my MLAG configuration.
2x CRS354-48G peer’d via QSFP 1-1 (01/02)
MLAG 2x2-4Port total LACP to adjacent switch (Procurve)
A third CRS354-48G (03) is connected to switch 354-01 via single QSFP 2-1
Video server connected to 354-03 while cameras are fed from Procurve.
I see Bulk of data streams going to 354-02, traversing the PEER to get to SW-01 to then traverse the standard QSFP link to 354-03.
Even though the most direct path is MLAG to 354-01 > QSFP > 354-03.
So while in your example two clients MLAG between the two switches directly may work as expected, it’s not always the case and PEER link capacity does need to be considered.
Device which uses LAG to connect to two different switches doesn’t know (nor care) about topology beyond its link peer(s) and selects one of physical links for Tx acording to selected Tx strategy (e.g. L2+L3 hash). If it selects “wrong” link, then of course switch has to pass that traffic to the other MLAG peer over peer link.
And vice versa: in your particular case CRS-01 will pass all traffic towards Procurve directly, none of that traffic will pass MLAG peer link (as long as direct link is alive). If you had a (normal) LAG link with two members between CRS-01 and Procurve, then CRS-01 might select different physical links for part of traffic (depending on Tx strategy).
Yeah I am aware an not surprised by the way my traffic is traversing the switches.
I was just adding insight to OP’s question about what traffic can travers the PEER interface. In a scenario where two clients are MLAG to the same pair of switches its unlikely to see much PEER link traffic, however in my scenario it totally needs to be included in the topology decision.
Fortunately for me 40Gb links between the CRS354s is completely future proof for my needs an I will never have to worry about link saturation.
