On the 7210 SAS-K 2F1C2T, 7210 SAS-K 2F6C4T, and 7210 SAS-K 3SFP+ 8C, an ingress QoS policy is applied to the aggregate traffic that is received through all the member ports of the LAG and mapped to that service entity (for example: access-uplink port). For example, if an ingress policy is configured with a queue shaper rate of PIR 100Mbps for an access-uplink LAG configured with two ports, then the queue shaper limits the traffic received through the two ports to a maximum of 100Mbps.
On the 7210 SAS-K 2F1C2T,7210 SAS-K 2F6C4T, and 7210 SAS-K 3SFP+ 8C, egress QoS policy parameters are applied to all the ports that are members of the LAG (all ports get the full SLA). For example, if an egress policy is configured with a queue shaper rate of PIR 100Mbps, and applied to an access-uplink LAG configured with two port members, then each port can send out 100 Mbps of traffic for a total of 200Mbps of traffic out of the LAG (assuming flows are distributed among the 2 ports). The advantage of this method over a scheme where the PIR is divided equally among all the member ports of the LAG is that a single flow can use the entire SLA. The disadvantage is that the overall SLA can be exceeded if the flows span multiple ports.