4A0-D01 Nokia Data Center Fabric Fundamentals Practice Questions

SR Linux allows routing protocols and static routes to be configured on a per-network-instance basis. To create a default route that is only visible within the 'customer-a' VRF, the administrator must configure the static route inside that specific network instance's context. This ensures the route does not leak into the global routing table or other VRFs unless explicitly configured to do so. The next-hop for this static route would be the firewall's IP address, which is available in the VRF's routing table due to the route leaking configuration.

The summary keyword provides a condensed, one-line-per-neighbor view of the BGP sessions. It is used for a quick health check to see which neighbors are up or down, how long they have been in that state, and the number of prefixes received. It does not show the actual prefixes themselves.

SR Linux uses a centralized logging and event handling system. To capture events from a specific application like BGP, you configure rules under /configure system logging. You can create a rule that matches on the 'bgp' application and directs those logs or events to a file, a remote syslog server, or the console.

The commit command in SR Linux applies the candidate configuration to the running configuration, making it active. However, it does not automatically save it to persist across reboots. The explicit command tools system configuration save must be executed to write the current running configuration to the startup configuration file.

Oversubscription ratio is calculated as the total bandwidth of server-facing ports (downlinks) divided by the total bandwidth of spine-facing ports (uplinks). A 3:1 ratio means the downlink bandwidth is three times larger than the uplink bandwidth. In the example diagram, total downlink bandwidth is 6 * 10G = 60G, and total uplink bandwidth is 2 * 10G = 20G. The ratio is 60:20, which simplifies to 3:1. This implies that if all servers transmit at full capacity simultaneously, contention and potential packet drops will occur at the uplinks.

Bidirectional Forwarding Detection (BFD) is a low-overhead, short-duration protocol designed to provide very fast failure detection for various routing protocols, including BGP. It operates independently of the routing protocol's own timers, allowing for sub-second detection of link or node failures, which then signals BGP to tear down the session and reconverge much faster than relying on BGP's slower keepalive and hold timers. STP is a Layer 2 protocol used to prevent loops and is not used in a Layer 3 leaf-spine fabric. LACP is used for link aggregation. OSPF is another routing protocol and would not be used to monitor an eBGP session.

In the SR Linux architecture, the Management Server is the central component that handles configuration transactions. When a configuration is committed via an interface like the CLI or gNMI, the Management Server validates it against the YANG models and then translates it into the protobuf format that the IDB uses. The IDB then notifies the relevant applications (like the BGP application) of the state change. The Application Manager is responsible for the lifecycle of applications, not configuration translation.

IP-VRF (Virtual Routing and Forwarding) is the standard technology used to create multiple, independent routing table instances within a single physical router. This allows for complete Layer 3 isolation between tenants, preventing traffic from one customer's network from being routed to another's. MAC-VRF is used for Layer 2 segmentation, while subinterfaces are used to logically divide a physical interface but do not provide routing table isolation on their own.

Zero Touch Provisioning (ZTP) is an automated process that enables a network device to provision itself without manual CLI commands. Typically, the device sends a DHCP request on its management port, and the DHCP server responds with options that provide the location of a bootstrap configuration script and the final software image, allowing the device to configure itself and join the network automatically.

In a leaf-spine (or Clos) topology, the rule is that leaf switches connect only to spine switches, and spine switches connect only to leaf switches. There are no direct connections between leaf switches or between spine switches. All traffic between leaves, including east-west traffic, traverses a spine switch, ensuring a consistent and predictable number of hops (always two: leaf-spine-leaf).