Routing

 View Only
last person joined: 2 days ago 

Ask questions and share experiences about ACX Series, CTP Series, MX Series, PTX Series, SSR Series, JRR Series, and all things routing, including portfolios and protocols.

Traffic shaping on MX lags

  • 1.  Traffic shaping on MX lags

    Posted 05-06-2021 06:34
    Hi, 

    I would need a clarification on how traffic shaping works on lags on MX. I understand we can use either scale mode to devide the configured shaping-rate (which apparently is default one) or replicate mode to just replicate it. I believe we use scale mode on our network since we have not configured explicitely the mode.
    But I am not sure about something: usually we apply a shaper via a traffic-control-profile on our interfaces ge or xe. When we are not limited in terms on bandwidth on the link we just configure a shaper to 1g or 10g depending on the interface type. Which I believe is ok.
    My question is about lags because I see that we configure a shaper to the maximum bandwidth when all links are up in the lag, for example 20g for 2x10G lag. But what happens in  case  one link goes down? Does the scale mode still allocate 20g to the remaining interface (which would mess the queuing I believe) or is it going to ignore the shaping-rate and allocate 10G to it? 
    Would it be better to configure the shaping with "percent 100" instead a value in gig to adapt it automatically to the available bandwidth on the lag? 
    In this case where we want to use all the bandwidth on the lag, is there a benefit to apply a shaper at all? or would applying  a scheduler map directly on the lag be good enough? (same question for regular ge/xe interfaces)

    Thanks in advance! 
    Aurélien