Wired Intelligent Edge

Hello, from what I've understood actually a maximum of 128 (Aruba 8400) and 32 (Aruba 8320) VSX LAGs (as known as Multi-Chassis LAGs) can be defined and operated per node (and so globally) on a VSX deployment.

 

Question: are those numbers - especially for a VSX depolyment with Aruba 8320 - fixed and immutable or is there a plan to enhance them with future ArubaOS-CX 10.01 (or 10.0n) software releases?

 

I'm facing the above question when considering a scenario with two Aruba 8320 JL479A (48 SFP+ and 6 QSFP+ interfaces), deployed using VSX: in that case the VSX can, under some scenarios, easily outreach 30 or more Multi-Chassis LAGs globally especially when the number of MC-LAG member interfaces is kept as low as possible (say one or two interfaces per node)...in that case the number of possible MC-LAGs using no more than one or two interfaces per node can scale rapidly (with respect to the scenario where MC-LAGs member interfaces number is pushed to its actual limit of 4 member interfaces per MC-LAG per node, 8 member interfaces per MC-LAG globally on the entire VSX).

 

Side note: normal (non MC-) LAG maximum number should remain 128 for both Aruba 8320 and Aruba 8400, VSX or Standalone they are deployed.

We are working on improvements to increase the number of VSX Lags for Aruba 8000 series platforms. I sent you a PM with more information.

Hello Scott, thanks for answering (It's Sunday!).

I answered you in detail via PM too.

My question is not due - as of now - to a real blocking setup issue: as I wrote above...we're going to deploy VSX [*] using 28 one-port MC-LAGs (VSX LAGs) per chassis...so just below the actual 32 interfaces limit although expecting a little more headroom would be reasonable, especially on Aruba 8320 and for a similar corner case.

Our setup timeline will be some date of August, probably 1st or 2nd week of August.

Will be interesting to see upcoming enhancements.

 

[*] briefly discussed in this thread.

Davide.