@nprofile1q... I'm now pondering Juniper, MX960 maybe? I wonder what we peaked at with them, maybe 300? (add in the other chassis original MX models and we might have been over 2000 across the company circa 2018)
@nprofile1q... why down to 6.8 not 5.8? some form of interconnect thing that means it's still faster in the pipeline, but not able to achieve the theoretical speedup?
@nprofile1q... I have a few friends that have played with using ssh instead of login on serial gettys, there's a bit of flag wrangling needed, but it does work and would give you that flexibility