Load Balancing Algorithms - CPU Affinity

[sdebarros]

Hi
Everyone.

I am having difficulty trying to figure this one out..  I have gone through the DOC's, but still does not make sense.

Let me know if I got this right..

The request comes into the SE and this request may be on lets say the 2nd core of the SE..  Does does it send it to a Backend server that is using that same core??  If so, how does it know what cores another VM is using??

Like I said maybe I am totally off on this one.

Thanks again.

[TronAr]

Hehe, I remember this stuff confused me for a while.

Hope this helps: It's not about other VM, but the cores at the SE. AVI uses a tuned S.O. in the SEs, and a "service pipeline" so to say that tries to fine tune to the limit. In so doing, a core on the SE might get a fixed server mapping if using this algorithm for a VIP. Mind you, this might be wrong, but at this level of details, the only ones that know are AVI people, and they are not listening here AFAIK.

HTH

 

[TronAr]

Look here, if you have not already:

https://avinetworks.com/docs/20.1/service-engine-group/ under "Hyper Threading modes" for some insight to SE core management...

[sdebarros]

Hi TronAr.

 

Thanks a ton for your help.  I also decided to send the question to the email in the course material Module 01..

The one that says "Questions After Class: [e-mail]"  Looks like no one reads that email either..

Really sad...