Manage link definitions in traffic shaping
To manage a link, the traffic shaper needs the following information:
The speed of the link in both the send and receive directions.
Whether the link is a WAN link or a LAN network.
A way of distinguishing link traffic from other traffic.
The direction in which traffic is flowing over the link.
Link Speed— Link speed always refers to the speed of the physical link. In the case of a WAN link, it is the speed of the WAN segment that terminates in the building with the NetScaler SD-WAN WANOP appliance. The speed of the other end of the link is not considered. For example, the following figure shows a network of four appliances. Each appliance has its incoming and outgoing bandwidths set to 95% of the speed of its own, local WAN segment, without regard to the speed of the remote endpoints.
Figure 1. Local bandwidth limits track local link speeds
The reason for setting the bandwidth limits to 95% of the link speed instead of 100% is to allow for link overhead (few links can carry data at 100% of their published speeds) and to ensure that the appliance is slightly slower than the link, so that it becomes a slight bottleneck. Traffic shaping is not effective unless the traffic shaper is the bottleneck in the connection.
Distinguishing different types of traffic—In each link definition, you must declare whether the definition applies to a WAN link or a LAN network.
The traffic shaper needs to know whether a packet is traveling on the WAN, and, if so, in which direction. To provide this information:
For simple inline deployments, you declare that one port of the accelerated bridge belongs to the WAN link and that the other port belongs to the LAN.
In other deployment modes, the appliance examines IP addresses, MAC addresses, VLANs, or WCCP service groups. (Note that testing for WCCP service groups is not yet supported.)
If a site has multiple WANs, the local link definitions must include rules that enable the appliance to distinguish traffic from different WANs.