CCIE ENTERPRISE INFRASTRUCTURE · RESILIENT ARCHITECTURE

Enterprise Routing & Resilient

Network Architecture

Module 3

RIP (RIPv2)

Programme Enterprise Routing & Resilient Network Architecture (ERRNA)

Target cert CCIE Enterprise Infrastructure v1.1 (lab) + CCNP ENCOR 350-401

Primary platform Cisco IOS XE — with IOS XR, NX-OS, FRRouting, VyOS, Linux, AWS, Azure, GCP

Format Module of a multi-module professional course manual

Consistency contract: device names, addressing, ASNs, areas, zones, communities and route tags are fixed in
Module 1 and reused unchanged across all modules.
 Module 3 — RIP (RIPv2)

Why RIP is here. RIP is not on the CCIE EI v1.1 blueprint (Module 1 §3.3). It earns its place as
the cleanest possible vehicle for the distance-vector ideas you must understand to reason about
EIGRP, redistribution loops and convergence: routing-by-rumour, hop-count, the timers, split-
horizon, poisoning, and count-to-infinity. Learn the behaviour here; do not over-invest in RIP itself for
the exam.
Continuity. The reference enterprise (Module 1) runs EIGRP/OSPF/IS-IS/BGP — not RIP. So this
module uses a dedicated, realistic legacy acquisition island that runs RIPv2 and later
redistributes into the enterprise at a branch boundary. Addressing stays within the Module 1 plan
(branch [Link]/16 ) plus an acquired classful block [Link]/16 chosen specifically to expose
auto-summary.

Platform: Cisco IOS XE primary; XR/NX-OS/FRR/VyOS/Linux equivalents in §3.18. No invented

commands.

Table of Contents — Module 3

1. Objectives and lab topology
2. Distance-vector and Bellman-Ford
3. RIPv2 on the wire: metric, messages, timers, tables
4. Loop-prevention mechanisms
5. Baseline configuration and command breakdown
6. Passive interfaces
7. Auto-summary and manual summarization (the discontiguous-network trap)
8. Default-route advertisement
9. Equal-cost paths and RIP route selection
10. Authentication
11. Timers and convergence tuning
12. Redistribution into and out of RIP
13. RIPng (IPv6) in brief
14. Verification and representative output
15. Common failures and systematic troubleshooting
16. Cross-platform quick reference
17. Guided labs 3A–3G
18. Independent / troubleshooting / design challenges
19. Review questions
20. Module close-out
21. Module 3 Answer Key (separated)

1. Objectives and lab topology

1.1 By the end of Module 3 you can
• Explain distance-vector operation, Bellman-Ford, and why RIP is "routing by rumour."
• Describe every RIP timer (update/invalid/holddown/flush) and what each protects against.
• Explain and demonstrate split horizon, route poisoning, poison reverse, triggered updates and count-
to-infinity.

Enterprise Routing & Resilient Network Architecture Module 3 · Page 2

 • Configure RIPv2 with version 2 , no auto-summary , passive interfaces, manual summarization,
authentication and a default route.
• Diagnose the discontiguous-network failure caused by auto-summary.
• Redistribute safely into and out of RIP with correct seed metrics.
• Troubleshoot a "silent neighbour" (RIP has no hellos, so this is different from OSPF/EIGRP).

1.2 Lab topology — the legacy RIP island

Why this design teaches well: [Link]/24 (on L1) and [Link]/24 (on L3) are two halves of
the classful [Link]/16 , separated by 10.40.x links. With auto-summary on (the default), both
edges advertise the summary [Link]/16 → a discontiguous network that black-holes (§7). LEG-
RTR-02 ↔ BRN-RTR-01 is the redistribution boundary used in §12 and again in Module 8.

Addressing for this module:

Element Value

LEG loopbacks [Link]/32 , .2 , .3

LEG-LANs (classful halves) [Link]/24 (L1), [Link]/24 (L3)

L1↔L2 [Link]/30 (L1 .1, L2 .2)

L2↔L3 [Link]/30 (L2 .5, L3 .6)

L2↔B1 (boundary) [Link]/30 (L2 .9, B1 .10)

RIPv2 carries masks, so /30 and /24 coexist fine once no auto-summary is set — the lab
deliberately starts wrong to make the lesson land.

2. Distance-vector and Bellman-Ford

2.1 The core idea
A distance-vector router knows only two things per destination: a distance (metric) and a vector (the
neighbour to send toward). It never sees the whole topology. It trusts what neighbours tell it — "routing
by rumour" — and re-advertises, adding its own cost.
Bellman-Ford, informally: each router's cost to a destination = min over neighbours of (cost
neighbour advertises + cost of the link to that neighbour). RIP's link cost is always 1 hop, so cost =
number of routers to cross.

Enterprise Routing & Resilient Network Architecture Module 3 · Page 3

 2.2 Contrast (forward reference)
Family Knows Examples Strength

Distance-vector neighbours' summaries only RIP, (EIGRP is "advanced DV") simple, low resource

Link-state full topology map OSPF, IS-IS fast, loop-resistant

Path-vector full AS path BGP policy, scale

The distance-vector blind spot — you can't see beyond your neighbour — is exactly why loops are
possible and why RIP needs the safeguards in §4. EIGRP (Module 4) keeps the DV model but adds the
feasibility condition to guarantee loop-freedom without those crude timers.

3. RIPv2 on the wire

Property RIPv2

Transport UDP, port 520

Destination multicast [Link] (RIPv1 used broadcast)

Metric hop count, max 15; 16 = infinity = unreachable

Update style entire table, every 30 s (periodic) + triggered on change

Masks carried (classless / VLSM) — RIPv1 did not

Neighbours no hello / no adjacency — routers just send and listen

Auth plaintext or MD5 (via key chains)

Control plane: build/maintain the RIP database from received updates; advertise the table every 30 s.
Data plane: ordinary CEF forwarding — RIP does nothing special to packets it forwards. No neighbour
state machine exists — a crucial troubleshooting difference (§15).

3.1 The four timers

Timer Default Meaning

Update 30 s how often the full table is advertised

Invalid 180 s if no update for a route in this time → mark metric 16 (possibly down)

Holddown 180 s after a route goes bad, ignore worse-or-equal news about it for this long (loop damper)

Flush 240 s remove the route entirely if still not refreshed

These come straight from timers basic 30 180 180 240 (§11). The large invalid/flush values are why
plain RIP convergence is slow — minutes, not seconds.

4. Loop-prevention mechanisms
Distance-vector's blind spot makes the count-to-infinity problem real: when a network dies, routers
can bounce a stale route back and forth, each incrementing the hop count, looping packets until the
metric crawls up to 16. RIP layers several defences:

Mechanism What it does Lab

Maximum hop count caps count-to-infinity; 16 = unreachable 3E

(16)

Split horizon never advertise a route back out the interface you learned it on 3E

Route poisoning when a route fails, advertise it with metric 16 (don't just stop advertising) 3E

Enterprise Routing & Resilient Network Architecture Module 3 · Page 4

 Mechanism What it does Lab

Poison reverse send the metric-16 poison back out the learning interface, overriding split horizon — 3E
explicit "don't use me"

Hold-down timer refuse worse/equal updates for a route just declared down, so stale rumours can't 3E/
revive it 3F

Triggered updates send an update immediately on a change instead of waiting up to 30 s 3F

Teaching payoff: these are the manual, timer-based versions of what EIGRP's DUAL does
deterministically and instantly. Understanding the crude version makes the elegant version (Module
4) obvious.

5. Baseline configuration and command breakdown

LEG-RTR-02#
router rip
version 2
no auto-summary
network [Link]
network [Link]
passive-interface default
no passive-interface GigabitEthernet0/0/0
no passive-interface GigabitEthernet0/0/1

Command breakdown

Command What / why / where / depends / if omitted / if misconfigured / verify

router rip Enters RIP config. Only one RIP process per router. Verify show ip protocols .

version 2 Forces RIPv2 (multicast, masks, auth). If omitted, IOS sends v1 and receives v1+v2 →
masks lost, neighbours mismatch. Always set it.

no auto-summary

Enterprise Routing & Resilient Network Architecture Module 3 · Page 5

 Command What / why / where / depends / if omitted / if misconfigured / verify
Disables automatic summarisation at classful boundaries. If omitted, discontiguous
[Link]/16 breaks (§7). Verify in show ip protocols ("automatic network
summarization is not in effect").

network [Link] Classful wildcard: enables RIP on every interface whose IP falls in [Link]/8 (sends/
receives updates + advertises those connected nets). Note you enter the classful network
even for subnets. If omitted, that interface neither runs RIP nor is advertised.

network [Link] Same, for the acquired block.

passive-interface Speak RIP only where intended (§6).

default + no passive-
interface …

The network gotcha: RIP's network statement is classful regardless of your subnetting — you
cannot scope it to a single subnet. To run RIP on some 10.x interfaces but not others, use passive-
interface (§6), not a narrower network .

6. Passive interfaces
A passive interface still advertises its subnet into RIP but stops sending updates out of it (it
keeps listening). Use it on LAN/host-facing and untrusted links so you don't leak routing to hosts or invite
rogue neighbours.

router rip
passive-interface default ! make everything passive…
no passive-interface GigabitEthernet0/0/0 ! …then re-activate only the inter-router links
no passive-interface GigabitEthernet0/0/1

Question Answer

What Suppresses RIP updates out an interface; subnet still advertised

Why Security + efficiency: no RIP toward hosts/users; no accidental peering

If omitted on a LAN Hosts see RIP multicast; a rogue device could inject routes

If applied to a transit link by mistake The neighbour stops learning routes through it (silent partial outage)

Verify show ip protocols → "Passive Interface(s)" list

7. Auto-summary and manual summarization

7.1 The discontiguous-network trap (auto-summary ON — the
default)
[Link]/24 (L1) and [Link]/24 (L3) are subnets of classful [Link]/16 , separated by 10.40.x
links. With auto-summary, each edge advertises the summary [Link]/16 across the 10.x
boundary. LEG-RTR-02 then receives [Link]/16 from both directions with equal hops and load-
balances into a black hole — half the destinations are unreachable.

Enterprise Routing & Resilient Network Architecture Module 3 · Page 6

 Symptom: intermittent reachability to 172.20.x hosts; show ip route shows [Link]/16 with two
equal next-hops. Fix: no auto-summary on every RIP router → the specific /24 s propagate and LPM
resolves correctly.

7.2 Manual summarization (interface-level, the right way)

Once auto-summary is off, summarise deliberately at a boundary:

interface GigabitEthernet0/0/0
ip summary-address rip [Link] [Link]

Question Answer

What Advertises a single [Link]/16 out this interface instead of the component /24 s

Why Smaller tables, hides internal flaps

Where The interface facing the rest of the network (e.g. toward BRN-RTR-01 ) — and only where the whole
block is reachable behind you

If Summarising at the wrong place re-creates the discontiguous black-hole

misconfigured

Verify show ip rip database (look for the summary), neighbour's show ip route rip

8. Default-route advertisement
router rip
default-information originate

Injects [Link]/0 into RIP so downstream routers get a gateway of last resort. Alternatively, redistribute
a static default:

ip route [Link] [Link] [Link]

router rip
redistribute static metric 1

Question Answer

What default-information originate advertises a RIP default to neighbours

Why Legacy island reaches the enterprise/Internet via one exit

Depends on A real path to send default traffic toward (don't originate a default you can't honour)

If misconfigured Black-hole: you advertise a default but have nowhere to send it

Verify downstream show ip route shows R* [Link]/0

Enterprise Routing & Resilient Network Architecture Module 3 · Page 7

 9. Equal-cost paths and RIP route selection
RIP selects the route with the lowest hop count. Equal-hop paths load-share — default 4, max 16 —
via:

router rip
maximum-paths 4

Because the only metric is hops, RIP cannot prefer a fast 2-hop path over a slow 1-hop path — a key
weakness motivating EIGRP's composite metric (Module 4). RIP also can't do unequal-cost sharing.

LEG-RTR-02# show ip route rip

R [Link]/24 [120/1] via [Link], 00:00:11, Gi0/0/1
R [Link]/32 [120/1] via [Link], 00:00:11, Gi0/0/1

[120/1] = AD 120, metric 1 hop.

10. Authentication
RIPv2 supports plaintext (weak) and MD5 (preferred) authentication via key chains. Without it, anyone
on a segment can inject routes.

key chain RIP-KEYS

key 1
key-string S3cur3-RIP-Key
!
interface GigabitEthernet0/0/0
ip rip authentication mode md5
ip rip authentication key-chain RIP-KEYS

Command breakdown

Command What / why / verify

key chain RIP-KEYS / key 1 / Defines the shared secret(s). Both ends need a matching key-id + string.
key-string …

ip rip authentication mode Use MD5 (omit for plaintext — don't). Must match on both ends or updates are
md5 silently dropped.

ip rip authentication key- Binds the key chain to the interface.

chain RIP-KEYS

If mismatched Updates ignored → routes silently disappear (looks like a dead neighbour). Verify
with debug ip rip (shows "ignored, authentication fail").

Key-chain rotation: accept-lifetime / send-lifetime under a key let you roll secrets without an
outage. Clocks must be in sync (NTP) for time-bound keys.

11. Timers and convergence tuning

router rip
timers basic 30 180 180 240
! update invalid holddown flush

Enterprise Routing & Resilient Network Architecture Module 3 · Page 8

 Lowering them speeds convergence but increases churn and risk; all routers in the domain must
agree or you get instability. RIP's floor on convergence (invalid 180 s) is why it's unsuitable for modern
networks. Triggered updates help on change, but recovery still waits on invalid/flush.

Question Answer

What Sets update/invalid/holddown/flush seconds

Why Trade convergence speed vs stability

Rule invalid and holddown should be ≥ several update intervals; flush > invalid

If mismatched across routers Premature flushing, route flap, loops

Verify show ip protocols (shows current timers and next-update countdown)

12. Redistribution into and out of RIP

This is where the legacy island meets the enterprise — and where seed metrics bite. RIP's metric is hops,
so any route entering RIP must be given a hop metric, or it is treated as unreachable.

12.1 Into RIP (enterprise → legacy)

router rip
redistribute ospf 1 metric 3
redistribute eigrp 100 metric 3
default-metric 3 ! fallback seed metric for all redistribution

Rule Why

Always specify metric or Without a seed metric, redistributed routes get an infinite metric and are not
default-metric advertised

Keep the seed modest (e.g. 1– Leaves headroom under the 15-hop ceiling for the legacy island to traverse
3)

Filter what you inject Don't dump the full enterprise table into a 15-hop-limited domain — summarise/
whitelist (route-map + prefix-list from Module 2 §18)

12.2 Out of RIP (legacy → enterprise)

Done on the enterprise side ( BRN-RTR-01 ) into EIGRP/OSPF — covered fully in Module 8, but the boundary
mindset starts here:

! On BRN-RTR-01 (preview):
router ospf 1
redistribute rip subnets metric-type 1 tag 999

• subnets is mandatory or only classful networks redistribute into OSPF.

• tag 999 marks RIP-origin routes so they can be filtered/denied if they try to loop back (Module 1 §8.6
loop-prevention rule).

12.3 The redistribution loop risk (two boundary routers)

If two routers join RIP and the enterprise, a route can leave RIP, traverse the enterprise, and re-enter RIP
with a different metric/AD — a feedback loop. Defences (built properly in Module 8): tag on export, deny
the tag on import, and adjust AD. The legacy island here uses a single boundary to keep Module 3
focused; the two-boundary loop is a Module 8 lab.

Enterprise Routing & Resilient Network Architecture Module 3 · Page 9

 13. RIPng (IPv6) in brief
IPv6 uses RIPng (UDP 521, multicast FF02::9). It is enabled per-interface, not via network :

ipv6 unicast-routing
ipv6 router rip LEGACY6
interface GigabitEthernet0/0/0
ipv6 rip LEGACY6 enable

Same distance-vector behaviour, 15-hop limit, split-horizon/poison. Authentication relies on IPsec rather
than RIPv2-style key chains. Verify: show ipv6 rip , show ipv6 route rip .

14. Verification and representative output

Command Shows

show ip route rip installed RIP routes ( R , [120/hops] )

show ip protocols version, networks, timers, passive ifaces, routing information sources (gateways), auto-
summary state, distance

show ip rip every prefix RIP knows + where learned + summaries

database

debug ip rip live updates sent/received, and auth failures / version mismatches

LEG-RTR-02# show ip protocols

Routing Protocol is "rip"
Sending updates every 30 seconds, next due in 8 seconds
Invalid after 180, hold down 180, flushed after 240
Default version control: send version 2, receive version 2
Automatic network summarization is not in effect
Maximum path: 4
Routing for Networks: [Link] [Link]
Passive Interface(s): GigabitEthernet0/0/2
Routing Information Sources:
Gateway Distance Last Update
[Link] 120 00:00:06
[Link] 120 00:00:11
Distance: (default is 120)

The "Routing Information Sources / Gateway" list is RIP's substitute for a neighbour table — if a
gateway you expect is missing or its "Last Update" is stale, that's your dead-neighbour clue (§15).

15. Common failures and systematic troubleshooting

RIP has no hellos and no adjacency, so "the neighbour is down" presents differently than in OSPF/
EIGRP — you infer it from missing updates, not from a state machine.

Symptom Likely cause Check / fix

No RIP routes learned at interface not in a network statement; or show ip protocols networks/passive;
all made passive; or version mismatch set version 2 both ends

Routes appear then vanish authentication mismatch, or invalid/flush debug ip rip for "authentication fail";
every few min expiring (updates not arriving) check the gateway's Last Update

auto-summary still on

Enterprise Routing & Resilient Network Architecture Module 3 · Page 10

 Symptom Likely cause Check / fix
Discontiguous network no auto-summary everywhere; verify in
black-hole show ip protocols

Slow recovery after a inherent invalid(180)/flush(240) timers accept it, or tune timers domain-wide; this
failure is why RIP is legacy

A 16+-hop destination hop-count diameter exceeded summarise, or the network is simply too
unreachable large for RIP

Default not received default-information originate missing, originate on the exit router; check passive/
or filtered/passive on the path ACL on the path

Loop / packets TTL- redistribution feedback or auto-summary tag+deny on redistribution; no auto-

expiring summary

Systematic process (RIP layer): 1. show ip protocols — is RIP running, v2, no auto-summary, right
networks, expected gateways present and fresh? 2. show ip route rip — are the prefixes present with
sane hop counts? 3. show ip rip database — learned vs advertised vs summarised? 4. debug ip rip
(brief) — updates flowing? auth/version errors? 5. Confirm passive-interface isn't silencing a transit link.
6. Validate the fix; re-check the gateway "Last Update" stays fresh.

16. Cross-platform quick reference

Task IOS XE IOS XR NX-OS ( feature FRRouting VyOS
rip )

Enable router rip / router rip / router rip router rip / set protocols rip
version 2 interface … TAG ; iface ip version 2
router rip TAG

Run on network [Link] interface … ip router rip network set protocols rip
iface (under router TAG (per iface) [Link]/8 or interface ethX
rip) network IFACE

No auto- no auto-summary n/a (classless) no auto- no auto- default off

summary summary summary (default
off)

Passive passive-interface passive under passive- passive- set … rip interface

… iface interface interface … ethX (none = passive
default model differs)

Summarize ip summary- route-policy ip summary- (offset/route- prefix-list/policy

address rip … address rip … map)

Auth (MD5) key chain + ip rip keychain key chain key chain set … rip interface
authentication … … authentication

Default default- default- default- default- set protocols rip

originate information information information information default-information
originate originate originate originate originate

Verify show ip route show rip show ip rip show ip rip show ip route rip
rip / show ip route (vtysh)
protocols

Linux: no native RIP daemon; run FRR ripd or BIRD. Clouds: RIP is not offered by AWS/Azure/GCP
— their managed routing is static/BGP only, so RIP simply doesn't apply there (noted for
completeness).

Enterprise Routing & Resilient Network Architecture Module 3 · Page 11

 17. Guided labs 3A–3G
Build on the §1.2 island (LEG-RTR-01/02/03 + BRN-RTR-01). FRR/VyOS substitutes work throughout.

Lab 3A — Multi-router RIPv2 baseline

Objective: three-router RIP domain with full reachability. Steps: address per §1.2; on each LEG router
router rip / version 2 / no auto-summary / network [Link] / network [Link] ; make LAN/
loopback-facing interfaces passive. Verify: show ip route rip on L2 shows L1's and L3's /24 s and
loopbacks at [120/1] / [120/2] ; ping 172.20.1.x ↔ 172.20.3.x . Inject: shut L1↔L2; watch the route go
invalid then flush (time it — note the minutes). Reset: no router rip .

Lab 3B — Authentication
Objective: secure L1↔L2 with MD5; prove mismatch fails closed. Steps: configure key chain + ip rip
authentication mode md5 + key-chain on both ends of L1↔L2. Verify: routes still learned; now change
the key-string on one side only → routes vanish; debug ip rip shows "authentication fail." Reset:
remove auth.

Lab 3C — Auto-summary trap & manual summarization

Objective: reproduce and fix the discontiguous black-hole. Steps: temporarily re-enable auto-summary
on all LEG routers; observe [Link]/16 learned from both directions on L2 and broken host
reachability. Then no auto-summary ; confirm /24 s reappear. Finally apply ip summary-address rip
[Link] [Link] on L2→B1 only. Verify: B1 sees one [Link]/16 ; intra-island still uses /24 s.
Reset: remove summary, keep no auto-summary .

Lab 3D — Default route

Objective: legacy island reaches "everything else" via B1. Steps: on the boundary path originate a
default ( default-information originate on the router with the real exit) or redistribute a static default
with a seed metric. Verify: L1 and L3 show R* [Link]/0 ; trace toward an off-island address exits via
L2→B1. Reset: remove default.

Lab 3E — Loop-prevention demonstration

Objective: see split-horizon, poisoning and poison-reverse in action. Steps: debug ip rip on L2; flap
L3's [Link]/24 ; observe the triggered poison (metric 16) and poison-reverse messages, and
that L2 will not re-learn the dead route from the other side. Then no ip split-horizon on a link and
watch a count-to-infinity creep toward 16. Verify: debug output matches the §4 sequence diagram.
Reset: restore split-horizon.

Lab 3F — Convergence analysis

Objective: quantify RIP's slow recovery and the effect of timers. Steps: with default timers, fail a link
and time until the route flushes and an alternate installs. Then set timers basic 10 30 30 40 on every
router and repeat. Verify: record both convergence times; note the trade-off (faster but chattier/
riskier). Discuss why this floor makes RIP unfit for the reference enterprise. Reset: restore default
timers.

Enterprise Routing & Resilient Network Architecture Module 3 · Page 12

 Lab 3G — Broken-neighbour troubleshooting
Objective: find why L3's routes stopped arriving at L2. Setup (instructor): plant one of: L2↔L3 link
made passive; version mismatch (L3 left at v1); auth mismatch; or network statement missing on L3.
Task: using only show ip protocols (gateway Last Update), show ip route rip , and brief debug ip rip ,
identify which fault is present and fix it. Verify: L3 reappears as a fresh gateway; routes restored.
Reset: clear the planted fault.

18. Challenges
18.1 Independent configuration challenge
On a clean four-router island deliver: RIPv2 everywhere with version 2 + no auto-summary ; all host/
loopback interfaces passive; MD5 auth on every inter-router link; a single [Link]/16 summary
advertised only toward BRN-RTR-01 ; a RIP default originated from the boundary; and maximum-paths 4 .
Provide the exact verification sequence proving each requirement. (Outline in §21.)

18.2 Troubleshooting challenge

A config is loaded with three planted faults: (a) one router still summarising automatically, (b) one inter-
router link silently passive, (c) one redistributed route entering RIP with no seed metric (so it never
advertises). Symptoms: partial 172.20.x reachability, a missing neighbour, and an enterprise prefix that
"should be in RIP but isn't." Find all three; for each give symptom → revealing command → root cause →
fix. (Answers in §21.)

18.3 Design exercise

The acquired company will be folded into the enterprise over six months. Argue, in a short design note,
why you would not extend RIP into the enterprise and instead (a) contain RIP behind a single tagged
redistribution boundary, (b) summarise the acquired space, and (c) migrate the island to EIGRP/OSPF.
Identify two concrete risks if a second RIP↔enterprise boundary were added without loop controls.
(Discussion in §21.)

19. Review questions

1. RIP has no hello protocol. How do you tell that a RIP "neighbour" has gone away?
2. Why must you enter a classful network statement, and how do you then run RIP on only some
subnets of that classful network?
3. Explain the discontiguous-network failure and the one command that prevents it.
4. Distinguish route poisoning from poison reverse, and state which one overrides split horizon.
5. A redistributed OSPF route does not appear in RIP at all. What is the most likely single cause?
6. Default RIP invalid and flush timers are 180 s and 240 s. What real-world consequence does this have,
and why is it acceptable to lower them only if done everywhere?
7. Two equal-hop RIP paths exist to a destination, one over a 1 Gbps link and one over a 10 Mbps link.
Which does RIP prefer and why is that a problem?
8. What does [120/2] mean in show ip route ?
9. Why is RIP unsuitable as the IGP for the Module 1 reference enterprise? Give two specific reasons.
10. You see a route's metric climbing 11 → 12 → 13 across successive updates. What is happening and
what caps it?
(Answers in §21.)

Enterprise Routing & Resilient Network Architecture Module 3 · Page 13

 20. Module close-out
20.1 Summary
RIP is a pure distance-vector protocol — routing by rumour over UDP/520 to [Link], hop-count metric
capped at 15 (16 = unreachable), full table every 30 s, no neighbour adjacency. Its loop defences (max-
hop, split horizon, poisoning, poison reverse, hold-down, triggered updates) and its slow timers (invalid
180, flush 240) are the whole lesson: they make the distance-vector failure modes — count-to-infinity,
discontiguous-network black-holes from auto-summary, slow convergence — concrete. RIPv2 adds
masks, multicast and MD5 auth. Redistribution requires a hop seed metric and tagging to stay loop-
free. Everything painful here is what EIGRP (Module 4) and the link-state protocols solve.

20.2 Key commands

router rip / version 2 / no auto-summary / network <classful>

passive-interface default / no passive-interface <if>
ip summary-address rip <net> <mask> (interface)
default-information originate
maximum-paths <n> timers basic 30 180 180 240
key chain NAME / key N / key-string X
ip rip authentication mode md5 / ip rip authentication key-chain NAME (interface)
redistribute <proto> metric <hops> / default-metric <hops>
show ip route rip | show ip protocols | show ip rip database | debug ip rip

20.3 Common mistakes

• Leaving auto-summary on → discontiguous black-hole.
• Forgetting version 2 → masks dropped, v1/v2 mismatch.
• Redistributing into RIP with no seed metric → routes silently not advertised.
• Making a transit link passive → silent partial outage.
• Mismatched MD5 keys → routes vanish, looks like a dead neighbour.
• Expecting fast convergence — RIP's timers forbid it.

20.4 Verification checklist

• [ ] show ip protocols : v2, auto-summary off, correct networks, expected gateways fresh.
• [ ] show ip route rip : all expected /24 s and loopbacks at sane hop counts.
• [ ] No [Link]/16 learned from two directions (auto-summary clean).
• [ ] Auth applied + matching on every inter-router link.
• [ ] Default present where expected; summary only at the intended boundary.
• [ ] Redistributed routes carry a seed metric and a tag.

20.5 Practical assignment

Stand up the §1.2 island, get it fully converged and authenticated, summarise toward BRN-RTR-01 ,
originate a default, and commit configs to Git with a README that lists each timer's value and what it
protects against. Keep this island — Module 8 redistributes it into the enterprise and adds a second
boundary to demonstrate (and fix) a redistribution loop.

Enterprise Routing & Resilient Network Architecture Module 3 · Page 14

 20.6 Troubleshooting assignment
Deliberately set the four timers inconsistently across the three LEG routers (e.g. one router at timers
basic 10 30 30 40 , others default). Document the instability you observe (premature flushing, flap), then
explain why timer agreement is mandatory in any distance-vector domain — a principle that recurs with
EIGRP K-values in Module 4.

---

21. Module 3 Answer Key

Stop. Attempt everything first.

21.1 Review-question answers

1. RIP has no adjacency, so you infer a lost neighbour from the Routing Information Sources /
Gateway list in show ip protocols — a missing gateway or a stale "Last Update," plus the affected
routes ageing out (invalid → flush). debug ip rip confirms whether updates are arriving.
2. network is matched against classful boundaries by design (legacy of RIPv1), so you specify the major
network even for subnets. To run RIP on only some subnets of that classful network, leave the network
statement broad and use passive-interface to silence the interfaces you don't want peering on
(subnets still get advertised; updates just aren't sent there).
3. Two subnets of one classful network are separated by a different major network; auto-summary
makes each side advertise the classful summary, so a middle router learns the summary from both
directions and black-holes half the hosts. no auto-summary prevents it.
4. Route poisoning = advertising a failed route with metric 16 (explicitly "dead") instead of silently
withdrawing it. Poison reverse = sending that metric-16 poison back out the interface the route was
learned on — it overrides split horizon.
5. No seed metric on the redistribute (and no default-metric ): RIP treats the route as infinite-metric
and never advertises it. Add metric <hops> or default-metric .
6. Recovery after a failure takes up to ~3–4 minutes (invalid 180 + flush 240), so RIP is unfit for services
needing fast failover. Lowering timers speeds this but every router must use identical timers, or
mismatched expiry causes premature flushing, flapping and potential loops.
7. RIP prefers whichever is fewer hops; if both are 1 hop it may even load-share — it has no concept of
bandwidth, so it can pick the 10 Mbps path over the 1 Gbps path. EIGRP's bandwidth/delay composite
metric solves this.
8. AD 120 (RIP) and metric 2 hops to that destination.
9. Any two of: max 15-hop diameter (too small for the enterprise); minutes-long convergence (invalid/
flush timers); hop-only metric ignores bandwidth; full-table 30 s updates don't scale; weaker security/
feature set vs OSPF/EIGRP/IS-IS.
10. Count-to-infinity — a stale route is being re-learned and re-incremented around a loop because split-
horizon/poison didn't fully prevent it; the max hop count of 16 (= unreachable) caps it and
eventually clears the route.

21.2 Configuration-challenge outline (§18.1)

router rip / version 2 / no auto-summary / network [Link] / network [Link] on all; passive-interface
default + selective no passive-interface on inter-router links; key chain + ip rip authentication mode
md5 + key-chain on every transit link; ip summary-address rip [Link] [Link] only on the
L2→B1 interface; default-information originate on the boundary router; maximum-paths 4 . Verify: show

Enterprise Routing & Resilient Network Architecture Module 3 · Page 15

 ip protocols (v2, no auto-summary, passive list, gateways), show ip rip database (summary present),
debug ip rip (auth OK), show ip route rip on B1 (single /16 + default).

21.3 Troubleshooting-challenge faults (§18.2)

• (a) Auto-summary on one router: [Link]/16 appears from an unexpected direction in show ip
route rip ; fix no auto-summary .
• (b) Passive transit link: a neighbour gateway missing from show ip protocols ; routes through it
absent; fix no passive-interface <transit> .
• (c) No seed metric: redistributed enterprise prefix absent from show ip rip database despite the
redistribute line; debug ip rip shows it never advertised; fix metric <hops> / default-metric .

21.4 Design-exercise discussion (§18.3)

Contain RIP behind one tagged redistribution boundary so its slow timers/limited diameter can't degrade
the enterprise; summarise [Link]/16 so legacy instability is hidden; migrate the island to EIGRP/
OSPF on a schedule, then remove RIP. Two risks of a second uncontrolled boundary: (1) redistribution
feedback loop — a route exits RIP, crosses the enterprise, re-enters RIP with different metric/AD; (2)
suboptimal/asymmetric routing from inconsistent metrics across the two boundaries. Both are
mitigated by tag-on-export + deny-tag-on-import + AD control (Module 8).

End of Module 3. Module 4 — EIGRP — keeps the distance-vector model but replaces RIP's timers and
crude loop defences with DUAL, the feasibility condition, a real composite metric, and sub-second
convergence.

Enterprise Routing & Resilient Network Architecture Module 3 · Page 16