Hello, Would folks be willing to share their MPLS Core QoS model? We use "R" cards for Core links leaving us with 8 Forwarding Classes (FC) & 8 Queues (1-1 relationship). We decouple the services from the FC. There's only 8 FC so each service can't get dedicated FCs. We dedicate the NC FC for the routing protocols leaving 7 Q.
What I'm really interested in is how people handle scheduling. There seems to be 2 things you can tune: -BW % allocation (transmit-rate) -Priority (low, medium-low, medium-high, high, strict-high) If you keep the same priority on each FC, each Q gets the minimum allocated BW assigned by the %. Excess is shared differently depending on card type or tuning. If you set different priority for each Q, the % doesn't seem to do much. Which model do you use? Do you use Priority? If not, how do you allocate BW to each Q? And finally, do you use WRED? In every Q or just some? Thanks, Serge _______________________________________________ juniper-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/juniper-nsp
