Static routing

Why it exists

Dynamic protocols need neighbours, timers and a shared metric. Many edges have one uplink and one set of destinations that never change. Writing the next hop once is cheaper than running a protocol for nothing.

Router requirements treat static routing as a first-class mechanism: a prefix, a next hop, and a configurable preference against dynamic sources. Floating statics are that preference used as a backup. RFC 1812 §7.4STATIC ROUTINGRFC 1812 · PROPOSED STANDARD · June 1995

A next hop written by hand

A router has no dynamic protocol for a remote prefix. The operator installs a static route; forwarding looks only at the destination address in the IPv4 header.

How it works

Forwarding reads the IPv4 Destination Address and selects the longest matching prefix. A static route is one more candidate in that table, authored by configuration rather than by a protocol. RFC 791 §3.1Internet Header FormatRFC 791 · INTERNET STANDARD · September 1981

When two routes cover the same prefix length, administrative preference decides. A floating static uses a worse preference so it stays unused until the preferred source withdraws. RFC 1812 §5.2.4.4Administrative PreferenceRFC 1812 · PROPOSED STANDARD · June 1995

A longer prefix always wins over a shorter one, regardless of source. A /32 static overrides a /16 from OSPF because longest match runs before distance is compared. RFC 1812 §2.2.5.2Classless Inter Domain Routing (CIDR)RFC 1812 · PROPOSED STANDARD · June 1995

On the wire

Constructed examples, encoded from the field table below them — not captured traffic.

Configure it

When it breaks

Symptom first, because that is what you have when it happens.

1. The static appears in the table and traceroute dies one hop past this router, or never leaves it.

   Narrow it down

   1. Confirm the next-hop address is reachable (ping / show to the next hop, not only to the destination).
   2. Check whether the next hop resolves through another static or a default that loops.
   3. On Ethernet, verify ARP for the next hop completes.

   Cause

   The route is installed against a next hop that cannot be used. Recursive resolution fails silently from the destination’s point of view.

   Fix

   Point the static at a connected next hop or a next hop that resolves through a working route. Remove recursive chains that depend on the route they are trying to support.

   RFC 1812 §5.2.4.3Next Hop AddressRFC 1812 · PROPOSED STANDARD · June 1995

2. The primary path is down and the backup static never takes over.

   Narrow it down

   1. Compare administrative distances of the primary source and the static for the same prefix.
   2. Confirm the primary route actually left the table — a stuck adjacency or redistributed copy can keep a “better” entry alive.
   3. Check tracking or BFD if the static is meant to withdraw when the uplink dies.

   Cause

   The floating static is not worse than what remains, or the primary source never withdrew. Distance only matters between candidates that are both present.

   Fix

   Raise the static’s distance above every remaining source of that prefix, or ensure the primary is truly gone before expecting failover.

   RFC 1812 §5.2.4.4Administrative PreferenceRFC 1812 · PROPOSED STANDARD · June 1995

3. A working IGP appears ignored for one destination while the rest of the table looks correct.

   Narrow it down

   1. Look for a longer prefix to that destination from any source, especially a host static.
   2. Compare prefix lengths before comparing administrative distances.
   3. Check whether a leftover troubleshooting static was left behind.

   Cause

   Longest match selected a more specific static. Distance never competed because the prefixes were different.

   Fix

   Remove or shorten the stray specific, or make it intentional and documented.

   RFC 1812 §2.2.5.2Classless Inter Domain Routing (CIDR)RFC 1812 · PROPOSED STANDARD · June 1995

4. A static via an Ethernet interface without a next-hop address generates ARP for remote destinations and saturates the segment.

   Narrow it down

   1. Inspect the static: exit interface only versus next-hop address.
   2. Check ARP tables for many incomplete or unexpected entries toward remote prefixes.
   3. Confirm the media is multi-access rather than true point-to-point.

   Cause

   Exit-interface statics treat the destination as on-link. On Ethernet that means ARP for every remote address.

   Fix

   Rewrite the static with an explicit next-hop IP on that interface. Reserve exit-interface form for true point-to-point links.

5. After advertising a summary, destinations inside the hole bounce until TTL expires instead of failing fast.

   Narrow it down

   1. Compare the advertised aggregate against prefixes actually held.
   2. Check for a discard/Null route covering the summary.
   3. Trace a hole address and see whether it follows a default back out.

   Cause

   The summary claims reachability the site does not have, and without a local discard the packet follows a less-specific default.

   Fix

   Install a discard route for the summary so uncovered addresses die here.

Design notes

Prefer next-hop address over exit-interface alone on multi-access media. An exit interface without a next hop forces the router to ARP for every destination as if it were on-link.

Use discard routes under summaries you originate. An accidental gap under an advertised aggregate should fail here with a counter, not bounce toward a default.

Static routes do not detect that the far side of a path is down unless something else withdraws them. BFD or a tracked object is what turns a static into a route that can leave the table when the path dies.

Misconceptions

“Static routes always beat dynamic protocols.”

Only when configured with a better administrative preference for the same prefix length. Floating statics deliberately lose until the preferred source is gone. RFC 1812 §5.2.4.4Administrative PreferenceRFC 1812 · PROPOSED STANDARD · June 1995

“A better administrative distance overrides a longer prefix.”

Longest match runs first. A /32 from anywhere beats a /8 with the best distance, because the two were never compared on preference. RFC 1812 §2.2.5.2Classless Inter Domain Routing (CIDR)RFC 1812 · PROPOSED STANDARD · June 1995

“If show ip route prints the static, forwarding works.”

The line proves configuration and RIB presence, not that the next hop resolves or that the far path is up. Recursive failure and silent drops both look like a healthy static entry.

More walkthroughs

Terms

Longest prefix match

The forwarding rule: among all routes containing a destination, the one fixing the most bits is used. Metric and administrative distance only compare routes to the same prefix and never override this.

Administrative distance

A local, configurable preference between routing protocols offering the same prefix. It never crosses the wire and runs only after longest-prefix match has chosen the prefix.

Floating static

A static route configured with a worse administrative preference than the preferred source for the same prefix length, so it stays unused until that source withdraws.

Check yourself

• Which IPv4 header field does a router use to select a static route?

  • Source Address
  • Destination Address
  • Protocol
  • TTL
• A /24 OSPF route and a /32 static both cover the destination. Which wins?

  • OSPF, if its administrative distance is better
  • The /32, because longest match runs before distance
  • Neither; the router load-shares
  • Whichever was configured last
• Why configure a backup static with administrative distance 210 when OSPF is 110?

  • So the static is preferred for load sharing
  • So the static stays unused until OSPF withdraws the same prefix
  • Because 210 is the default for all statics
  • To make the static more specific
• OSPF and a floating static both offer 10.9.0.0/16. OSPF is up. Which is used?

  • The static, because statics always win
  • OSPF, because its administrative distance is better
  • Both, per-packet
  • Neither until you clear the RIB
• A static via 10.255.0.1 is in the table but packets vanish. What should you check next?

  • Only the destination host’s firewall
  • Whether 10.255.0.1 itself resolves to a usable forwarding path
  • Whether the TTL is too low on the host
  • Whether OSPF hello timers match
• What is a discard (Null0) static under a summary for?

  • Load sharing across uplinks
  • Failing locally for addresses the summary claims but the site does not hold
  • Preferring statics over OSPF
  • Disabling ICMP
• Why do connected routes usually beat statics to the same prefix?

  • Connected routes are always more specific
  • Platforms assign connected a better administrative preference than static
  • Statics cannot cover connected prefixes
  • RFC 791 requires it
• A static uses only an Ethernet exit interface, no next-hop IP. What goes wrong?

  • Nothing; that is the preferred form
  • The router ARPs for remote destinations as if they were on-link
  • The route cannot enter the RIB
  • OSPF immediately withdraws it

Related

• Subnetting

  Divides an address space by moving a boundary through the bits, and decides which route wins by how many of those bits a prefix pins down.

• OSPFv2

  Each router floods a description of its links, then independently computes the same shortest-path tree — so the forwarding tables agree without exchanging routes.

• Redistribution

  Moves routes from one protocol into another, discarding the metric and the loop prevention that made them trustworthy.