CompTIA Review

Network+ · Chapter 10

Routing Protocols

Distance-vector, link-state, hybrid, and path-vector protocols; administrative distance; metrics; RIP versions; loops; and static vs dynamic routing.

Routing protocol types

TypeProtocolsHow it learns routes
Distance-vectorRIP, RIPv2, IGRPPasses its routing table to neighbors. No full topology map.
Link-stateOSPF, IS-ISLearns every link and builds the same topology map, then runs Dijkstra.
HybridEIGRPDistance-vector behavior with link-state features.
Path-vectorBGPRoutes between autonomous systems, tracking the AS path.

RIP has no topology table. It only trusts what its neighbors tell it.

Administrative distance (AD)Often missed

AD ranks how much the router trusts each route source. When two sources offer the same route, the lowest AD is installed. Metrics only break ties between routes from the same protocol.

SourceAD
Connected interface0
Static route1
EIGRP (internal)90
IGRP100
OSPF110
RIP120
External EIGRP170
Unknown255 (never used)
  • When a router receives two updates for the same remote network, AD is the first thing it checks. The route with the lower AD goes into the routing table.
  • Internal EIGRP routes carry AD 90. External EIGRP routes, redistributed into EIGRP from another source, carry AD 170.
  • AD 255 means the source is unknown. The router never uses a route with AD 255.

A static default route (AD 1) beats a RIP default route (AD 120), even if RIP has a lower metric.

Real world: a wise network engineer does not change the default administrative distance values for protocols. Leave AD at its defaults so every router makes the same, predictable choice.

Real world: watching the network with SNMP traps

Leaving AD alone does not mean ignoring routing. A senior engineer managed his team of network engineers with SNMP traps, so problems reached them as they happened. An SNMP (Simple Network Management Protocol) trap is an unsolicited, real-time alert that an SNMP-enabled device (the agent) sends to a central network management system (NMS) to report a critical event or status change, without waiting to be polled.

  • Transport: traps use UDP port 162. Polling, such as GET requests, uses UDP port 161.
  • Unsolicited alerts: devices push a message only when a trigger happens, which cuts the overhead of constant polling.
  • OID (Object Identifier): a dotted-decimal string that uniquely identifies the event or object.
  • Varbinds (variable bindings): key-value pairs with context, such as sysUpTime, snmpTrapOID, interface indexes, or error descriptions.
  • Timestamp and agent address: when the event happened and which device sent it.
Use caseExample traps
Interface stateslinkDown or linkUp when a physical or logical interface flaps
Hardware and environmentPower supply loss, fan failure, high temperature, memory exhaustion
Security and accessAuthentication failures, wrong community string attempts, firewall security triggers
Routing and protocol changesBGP session drops, EGP neighbor loss
FeatureSNMP trapSNMP INFORM
Acknowledged?No. Fire and forget; the agent never knows if the NMS got it.Yes. The NMS must send an acknowledgment.
If the message is lostIt is gone.The agent retries until it gets an ACK.
Best forHigh-volume, lower-priority alertsCritical enterprise infrastructure
  • Upgrade to SNMPv3. SNMPv1 and SNMPv2c send community strings (public, private) in clear text; SNMPv3 adds authentication and encryption.
  • Compile MIBs on the NMS. Load the vendor MIBs (for example Cisco or Juniper) into the management platform, such as PRTG, LogicMonitor, or Netdata, so raw OIDs show as readable text.
  • Rate-limit and filter. Prevent trap storms during a major outage by filtering redundant or low-priority traps, and pair traps with periodic polling for complete visibility.
  • Use Cisco EEM (Embedded Event Manager) to script custom triggers that turn CLI events or log messages into tailored SNMP notifications.

Exam memory hook: traps go to UDP 162, polls go to UDP 161. A trap is not acknowledged; an INFORM is. For more, see Cisco's SNMP trap configuration guide or LogicMonitor's SNMP trap overview.

Metrics

  • RIP uses hop count. 1 to 15 is usable, and 16 or more is unreachable.
  • OSPF uses cost: reference bandwidth (100 Mbps) divided by interface bandwidth. Faster links cost less.
  • EIGRP uses a composite metric of delay, bandwidth, reliability, and load, weighted by K values.

RIPv1 vs RIPv2Often missed

FeatureRIPv1RIPv2
Update methodBroadcastMulticast 224.0.0.9
IntervalEvery 30 secondsEvery 30 seconds
Administrative distance120120
Subnet masks in updatesNoYes
Networks supportedClassful onlyClassless and discontiguous
AuthenticationNoYes

RIPv2 multicasts. It does not broadcast.

Routing loops

A routing loop sends packets around a set of routers without ever reaching the destination. Each hop lowers the TTL, and the packet is dropped when TTL reaches 0.

TTL expiry on its own does not prove a routing loop; it can also mean the destination is too many hops away. The loop itself is packets circulating between routers and never arriving.

Split horizon

Split horizon is a loop-prevention rule used by distance-vector protocols such as RIP and IGRP. A router does not advertise a route back out the interface it learned that route on.

  • Example: Router B learns 10.0.0.0 from Router A on its link to A. B does not send that route back to A on that link.
  • B still advertises the route out its other interfaces.
  • Without split horizon, A and B can pass the same route back and forth, and the packets loop.

Split horizon blocks one route on one interface. It does not stop every update, and it does not stop broadcast storms on switches.

Static, default, and dynamic routing

  • Dynamic routing reroutes around failures with no administrator action.
  • Static routing is typed in by an administrator and stays until someone changes it.
  • Default routes also need an administrator to change them if they fail.
  • Dynamic routing costs more RAM, CPU, and bandwidth, but it gives resiliency.

IPv6 discovery and transition

  • NDP (Neighbor Discovery Protocol) discovers neighboring devices on IPv6 networks.
  • Router advertisements (RA) announce the gateway so hosts can find it.
  • Dual stack runs IPv4 and IPv6 side by side.
  • Tunneling carries IPv6 inside IPv4. Teredo and ISATAP are tunneling methods.

Interactive visuals

This chapter has interactive visuals for every review concept. Open them at /network-plus/ch10.