An uplink is the final network segment that connects a local site to a network. You can define multiple uplinks for a single network, giving the system flexibility to choose the best path for traffic. The SteelHead monitors uplink status and selects the most appropriate path for each packet, which helps optimize network usage and control. Uplinks are essential for building your network topology.

Uplinks are also important for QoS, as they specify available bandwidth for remote sites. This bandwidth is used by RiOS to calculate the bottleneck bandwidth across the full path.

Each uplink can be defined by its outgoing interface and, optionally, a next-hop gateway IP. You can assign different DSCP marks to flows on each uplink, allowing routers to prioritize traffic based on these marks.

The probing frequency of uplinks affects hybrid network scalability and performance. SteelHeads probe more frequently when a site has traffic and back off when there's no activity. By default, uplinks are probed every 2 seconds if traffic is present. If not, probing slows to a default Max Backoff Interval of 1800 seconds, which you can adjust per your network needs.

If probe responses are not received within the timeout period, the system marks the probe as lost. After missing a set number of probes (defined by a threshold), the uplink is considered down, and an alert is triggered. The SteelHead automatically reroutes traffic to another available uplink. Once the failed uplink recovers, traffic is redirected back to it.

Path selection only uses local uplinks. To simplify configuration across similar sites, you can use site connectivity templates—predefined sets of uplinks that can be cloned and reused at multiple locations. This is helpful for branch offices or other sites with the same uplink structure. For details about site templates, see About site connectivity templates.

For detailed information about defining tunneled uplinks, see the SteelHead User Guide.