Routing Basics and Longest Prefix Match

What Routing Does

Routing moves packets between IP networks. Before sending, a host compares the destination against its own address and mask. If the destination is in the same subnet, the host delivers it directly on the LAN using Layer 2. If the destination is remote, the host forwards the packet to its default gateway, the router that knows how to reach other networks.

Local or Remote Decision

The host ANDs its address and the destination against its mask; matching network portions mean local delivery.

Longest Prefix Match

When several routes could carry a packet, the router picks the most specific one, meaning the route with the longest prefix length. A /24 is more specific than a /16, and a /32 host route is the most specific of all.

With every route present, traffic to 10.1.2.55 takes 10.1.2.0/24 because it is the longest matching prefix. The default route is used only when nothing more specific matches. Administrative distance breaks ties between equally specific routes from different sources (a static route, AD 1, beats an OSPF route, AD 110), and metric breaks ties within a single protocol.

Routing Troubleshooting

PBQ-Style Thinking

Scenario: a router holds routes for 10.0.0.0/8 via Router A, 10.2.0.0/16 via Router B, and 10.2.5.0/24 via Router C. A packet to 10.2.5.99 goes to Router C, 10.2.9.99 goes to Router B, and 10.9.9.9 goes to Router A. Walk each destination against the table and pick the longest match; this is the single most common routing PBQ pattern.

Scenario: a branch subnet 192.168.40.0/24 can initiate traffic to a data center, but servers cannot reply. The branch edge has a route, but the data center core lacks a return route to 192.168.40.0/24. Traffic flows out and dies on the way back. Add or advertise the return route and verify firewalls permit the reply.

Routing Protocol Context

N10-009 wants conceptual recognition, not vendor configuration.

The selection order to recite is: longest prefix match first, then administrative distance to compare route sources, then metric within a protocol. Master that order and most routing-table questions resolve themselves.

Administrative Distance Versus Metric

Candidates often blur these two tie-breakers, but they operate at different stages. Administrative distance (AD) ranks the trustworthiness of route sources and only matters when two routes to the same prefix come from different protocols. A directly connected route has AD 0, a static route AD 1, EIGRP 90, OSPF 110, and RIP 120, so a static route to a destination will be installed over an OSPF route to the same destination even though OSPF may know a technically shorter path. Metric only breaks ties inside one protocol, choosing the lower-cost path among routes the same protocol learned.

The exam may show a static and a dynamic route competing; the answer hinges on AD, not metric. It may instead show two OSPF paths with different costs; there the lower metric wins.

Default Gateway Versus Default Route

These sound identical but live on different devices. A default gateway is configured on an end host and names the single router the host uses for any off-subnet destination. A default route (0.0.0.0/0 or ::/0) lives in a router's table and tells that router where to send packets it has no specific route for. A host with a wrong default gateway cannot reach remote networks even if every router is perfectly configured; a router missing a default route may drop traffic to unknown destinations even though hosts are configured correctly. Read each scenario to determine which device owns the misconfiguration.

Return Paths and Asymmetry

The hardest routing PBQs are reachability problems where traffic flows one direction only. IP routing is hop-by-hop and independent in each direction, so a packet can reach a destination while the reply has no valid route home. The branch-to-data-center example earlier is the canonical case: outbound works because the path is fully routed, but the reply dies because the far side never learned a route back to the source subnet.

Always verify both a forward and a return route exist, and remember that stateful firewalls add a second requirement, the reply must traverse the same firewall that saw the request, or the connection state will not match and the firewall will silently drop it.