Routers and Routing Basics CCNA 2

1
Routers and Routing Basics CCNA 2
Chapter 7
2
Distance Vector Routing Protocols

• Using Distance Vector Routing Protocols
• Review of Distance Vector Operation in a Stable
  Network
• Route Poisoning
• Problem Counting to Infinity
• Loop-Prevention Features
• Summarizing Loop Avoidance
• Routing Information Protocol
• Configuring RIP Versions 1 and 2
• RIP Verification and Troubleshooting
• Choosing the Best Route Among the Possible
  Routes
• Integrating Static Routes with RIP
• Classful and Classless Routing Protocols,
  Routing, and Addressing
• Summary

3
Routing Loops the Fee for Simplicity

• In 1980s a typical WAN link to a remote site
  might have been only 56 kbps
• As a result, the designers of the first distance
  vector protocols had to keep them simple
• The simplicity of distance vector protocols
  introduced the possibility of routing loops (same
  single packet ends up back at the same routers
  over and over again)
• The looping packets could easily congest the
  network and make it unusable
• Routing loops must be avoided as much as possible

4
Distance Vector Operation in a Stable Network
Normal Steady-State RIP Operations

• 1. R2 considers itself to have a 0-hop route for
  subnet 172.30.22.0/24, so in the routing update
  sent by R2, R2 advertises a metric 1 (hop count
  1) route.
• 2. R1 receives the update, and because R1 has
  learned of no other possible routes to
  172.30.22.0, this route must be R1s best route
  to the subnet.
• 3. R1 adds the subnet to its routing table,
  listing it as a RIP-learned route.
• 4. For the learned route, R1 uses an outgoing
  interface of S0/0, because R1 received R2s
  routing update on R1s S0/0 interface.
• 5. For the learned route, R1 uses a next-hop
  router of 172.30.1.2, because R1 learned the
  route from a RIP update whose source IP address
  was 172.30.1

At the end of this process, R1 has learned a new
route. The rest of the RIP-learned routes in this
example follow the same process.
5
Distance Vector Operation in a Stable Network
(Continued)

• MetricRIP uses hop count for the metric. RIP
  routers add 1 to the metric before advertising
  the route.
• PeriodicThe hourglass icons represent the fact
  that the updates repeat on a regular cycle. RIP
  uses a 30-second update interval by default.
• Full updatesThe routers send full updates, every
  time, instead of just sending new or changed
  routing information. (The term partial update
  refers to routing updates that include only
  changed information.)
• Full updates limited by split horizon rulesThe
  routing protocol omits some routes from the
  periodic full updates due to split horizon rules.

6
Route Poisoning

• When a route fails, distance vector routing
  protocols risk causing routing loops until every
  router in the internetwork knows and believes
  that the original route has failed.
• As a result, distance vector protocols need to
  have a way to specifically identify which routes
  have failed.
• Distance vector protocols spread the bad news
  about a route failure by poisoning the route.
• Route poisoning refers to the practice of
  advertising a route, but with a special metric
  value called infinity.
• Simply put, routers consider routes advertised
  with an infinite metric to have failed. Each
  distance vector routing protocol uses the concept
  of an actual metric value that represents
  infinity.
• RIP defines infinity as 16.

7
Route Poisoning (Continued)

• 1. R2s FA0/1 interface fails.
• 2. R2 removes its connected route for
  172.30.22.0/24 from its routing table.
• 3. R2 advertises 172.30.22.0 with an infinite
  metric, which for RIP is metric 16.
• 4. R1 keeps the route in its routing table, with
  an infinite metric, as part of the loop-avoidance
  process.

8
Counting to Infinity

• Distance vector routing protocols risk causing
  routing loops between the time at which the first
  router realizes a route has failed until all the
  routers know that the route has failed.
• That problem, called counting to infinity, causes
  two other related problems
• 1. Packets may loop around the internetwork
  while the routers count to infinity, with the
  bandwidth consumed by the looping packets
  crippling an internetwork.
• 2. The counting-to-infinity process may take
  several minutes, meaning that the looping could
  cause users to believe that the network has
  failed.

9
Counting to Infinity (Continued)

• 1. R2s FA0/1 interface fails, so R2 removes its
  connected route for 172.30.22.0/24 from it
  routing table.
• 2. R2 sends a poisoned route advertisement
  (metric 16 for RIP) to R1, but at about the same
  time, R1s periodic update timer expires, so R1
  sends its regular update, including an
  advertisement of 172.30.22.0, metric 2.
• 3. R2 hears about the metric 2 route to reach
  172.30.22.0 from R1. Because R2 no longer has a
  route for subnet 172.30.22.0, R2 adds the two-hop
  route to its routing table, next-hop router R1.
• 4. At about the same time as Step 3, R1 receives
  the update from R2, telling R1 that its former
  route to 172.30.22.0, through R1, has failed. As
  a result, R1 changes its routing table to list a
  metric of 16 for the route to 172.30.22.0.

R2 Incorrectly Believes R1 has a Route to
172.16.22.0/24
10
Counting to Infinity (Continued)
R1 and R2 Count to Infinity

• 1. Both R1s and R2s update timers expire at
  about the same time. R1 advertises a poison
  (metric 16) route, and R2 advertises a metric 3
  route. (Remember, RIP routers add 1 to the
• metric before advertising the route.)
• 2. R2 receives R1s update, so R2 changes its
  route for 172.30.22.0 to use a metric of 16.
• 3. At about the same time as Step 2, R1 receives
  R2s update, so R1 changes its route for
  172.30.22.0 to use a metric of 3.

11
Loop-Prevention Features Split Horizon

• Split horizon is defined as follows
• In routing updates sent out interface X,
• do not include routing information about routes
• that refer to interface X as the outgoing
  interface.

12
Loop-Prevention Features Split Horizon
(Continued)

• 1. R1 sends its normal periodic full update,
  which, due to split horizon rules, includes only
  one route.
• 2. R2 sends its normal periodic full update,
  which, due to split horizon rules, includes only
  two routes.
• 3. R2s FA0/1 interface fails.
• 4. R2 removes its connected route for
  172.30.22.0/24 from its routing table.
• 5. R2 advertises 172.30.22.0 with an infinite
  metric, which for RIP is metric 16.
• 6. R1 temporarily keeps the route for 172.30.22.0
  in its routing table, later removing the route
  from the routing table.
• 7. In its next regular update, R1, due to split
  horizon, still does not advertise the route for
• 172.30.22.0.

The Effects of Split Horizon Without Poison
Reverse
13
Poison Reverse and Triggered Updates

• Distance vector protocols can attack the
  counting-to-infinity
• problem when reacting to failed routes by
  ensuring that
• every router learns that the route has failed,
  through every
• means possible, as quickly as possible
• Triggered update When a route fails, do not
  wait for the next periodic update. Instead, send
  an immediate triggered update listing the
  poisoned route.
• Poison reverse When learning of a failed
  route, suspend split horizon rules for that
  route, and advertise a poisoned route.

14
Poison Reverse and Triggered Updates(Continued)

• 1. R2s FA0/1 interface fails.
• 2. R2 immediately sends a triggered partial
  update with only the changed informationin
• this case, a poison route for 172.30.22.0.
• 3. R1 responds by changing its routing table and
  sending back an immediate (triggered) partial
  update, listing only 172.30.22.0 with an infinite
  metric (metric 16). This is a poison
• reverse route.
• 4. On R2s next regular periodic update, R2
  advertises all the typical routes, including the
• poison route for 172.30.22.0, for a time.
• 5. On R1s next regular periodic update, R1
  advertises all the typical routes, including the
• poison reverse route for 172.30.22.0, for a time.

R2 Sending a Triggered Update, with R1
Advertising a Poison Reverse Route
15
Loops in Redundant Networks

• Split horizon prevents the counting-to-infinity
  problem from
• occurring between two routers.
• However, with redundant paths in an
  internetwork, which is true of most internetworks
  today, split horizon alone does not always
  prevent counting to infinity.

16
Loops in Redundant Networks (Continued)
Periodic Updates in a Stable Triangle
Internetwork

• 1. R2 advertises a metric 1 route for 172.30.22.0
  in its updates to both R1 and R3.
• 2. R1 advertises a metric 2 route for 172.30.22.0
  to R3, while R3 advertises a metric 2 route for
  172.30.22.0 to R2.
• 3. Both R1 and R3 add the metric 1 route, learned
  directly from R2, to their routing tables, and
  ignore the two-hop routes they learn from each
  other. For example, R1 places a route
  172.30.22.0, using outgoing interface S0/0,
  next-hop router 172.30.1.2 (R2), in its routing
  table.

17
Loops in Redundant Networks (Continued)
Counting to Infinity in a Redundant
Internetwork, Part 1

• 1. R2s FA0/1 interface fails.
• 2. R2 immediately sends triggered partial updates
  poisoning the route for 172.30.22.0. R2 sends the
  updates out all still-working interfaces.
• 3. R3 receives R2s triggered update that poisons
  the route for 172.30.22.0, so R3 updates its
  routing table to list metric 16 for this route.
• 4. Before the update described in
• Step 2 arrives at R1, R1 sends its normal
  periodic update to R3, listing 172.30.22.0,
  metric 2, as normal. (Figure omits some of what
  would be in R1s periodic update to reduce
  clutter.)

5. R1 receives R2s triggered update (described
at Step 2) that poisons the route for
172.30.22.0, so R1 updates it routing table to
list metric 16 for this route. 6. R3 receives
the periodic update sent by R1 (described at Step
4), listing a metric 2 route for 172.30.22.0. As
a result, R3 updates its routing table to list a
metric 2 route, through R1 as the next-hop
router, with outgoing interface S0/0. At this
point, R3 has an incorrect metric 2 route for
172.30.22.0, pointing back to R1.
18
Loops in Redundant Networks (Continued)
Counting to Infinity in a Redundant
Internetwork, Part 2

• 7. R1 sends its next periodic update to R3, with
  poisoned route 172.30.22.0, metric 16.
• 8. Before the update described in Step 7 arrives
  at R3, R3 sends its next periodic update toR2,
  listing a metric 3 route for 172.30.22.0.
• 9. R3 receives R1s periodic update from R1 (as
  described in Step 7), and R3 changes its route
  for 172.30.22.0 to list an infinite metric.
• 10. R2 receives R3s periodic update (as
  described in Step 8), so R2 adds a metric 3 route
  for 172.30.22.0 to its routing table, listing R3
  as the next-hop router, with outgoing interface
• S0/1/

19
The Holddown Process and Holddown Timer

• Distance vector protocols use holddown to
  specifically
• attack the loops created by the
  counting-to-infinity
• problems that occur in redundant internetworks
• The term holddown gives a hint as to its meaning
• After the route is considered to be down, hold
  the route in a down state for a while to give the
  routers time to make sure every router knows that
  the route has failed.
• The holddown process tells a router to ignore new
  information about
• the failed route, for a time period called the
  holddown time, as counted
• using the holddown timer.

20
Using Holddown to Prevent Counting to Infinity

• 1. R2s FA0/1 interface fails.
• 2. R2 immediately sends triggered partial
  updates, poisoning the route for 172.30.22.0. R2
  sends the updates out all still-working
  interfaces.
• 3. R3 receives R2s triggered update that poisons
  the route for 172.30.22.0, so R3 updates its
  routing table to list metric 16 for this route.
  R3 also puts the route for 172.30.22.0 in
  holddown and starts the holddown timer (180
  seconds by default with RIP) for the route.
• 4. Before the update described in Step 2 arrives
  at R1, R1 sends its normal periodic update toR3,
  listing 172.30.22.0, metric 2, as normal. (Note
  that Figure 7-10 omits some details in R1s
  periodic update to reduce clutter.)
• 5. R1 receives R2s triggered update (described
  in Step 2) that poisons the route for
• 172.30.22.0, so R1 updates its routing table to
  list metric 16 for this route.

6. R3 receives the update from R1 (Step 4),
listing a metric 2 route for 172.30.22.0. Because
R3 has placed this route in a holddown state, and
this new metric 2 route was learned from a
different router (R1) than the original router
(R2), R3 ignores the new routing information.
21
Summarizing Loop Avoidance

• During periods of stability, routers send
  periodic full updates.
• The updates list all known routes except the
  routes omitted due to split horizon rules.
• When changes occur that cause a route to fail,
• routers react by sending triggered partial
  updates with poisoned routes.
• Routers also suspend split horizon rules for that
  route
• advertising a poison reverse route back toward
  the router from which the failed route was
  learned
• All routers place a route in holddown state and
  start a holddown timer for that route after
  learning that the route has failed.
• The router ignores all new information about that
  route until the holddown timer expires, unless
  that information comes from the same router that
  originally advertised the good route to that
  subnet.

22
Distance Vector Loop Avoidance Terminology
23
Routing Information Protocol

• The first IP networks used RIP Version 1 (V1)
  because it was the
• first and only IP routing protocol early in the
  history of TCP/IP.
• As time went on, routers became more affordable,
  with
• faster CPUs, more memory, and faster links, all
  of which allowed the
• development of more advanced routing algorithms
  and routing
• protocols, such as OSPF and EIGRP.
• Around the same time, other developers enhanced
  the RIP protocol
• standard, calling the new standard RIP Version 2
  (V2).
• RIP V2 does not completely change RIP V1, but
  rather adds some advanced features.

24
Comparing RIP Version 1 and 2 Features
25
Configuring RIP V1

• RIP V1 configuration requires two configuration
  commands
• - router rip
• - network classful-network-number
• The router rip command moves the user from global
  
• configuration mode to RIP configuration mode, and
  the
• network command tells the router on which
  interfaces to
• start using RIP.

26
Configuring RIP V1 (Continued)

• Configuring RIP on All Interfaces on R1

27
Configuring RIP V1 (Continued)

• When a routers RIP configuration matches an
  interface,
• Cisco IOS starts the following process
• 1. Sends RIP updates out the interface.
• 2. Listens for RIP updates coming in that
  interface from some other router.
• 3. Advertises the subnet attached to the
  interface.

28
Configuring RIP V2

• 1. To configure RIP V2 in internetworks that use
  RIP V2 only, simply add the version 2 command
  under router rip.
• 2. After they are configured, the routers send
  only V2 updates
• And process only received V2 updates.
• 3. At that point, the core features of RIP V2,
  such as sending
• masks in routing updates occur.
• 4. Optional RIP V2 features, such as
  authentication, this requires additional
  configuration.

29
Using Both RIP V1 and V2

• 1. In some cases, an internetwork may need to use
  both RIP versions.
• (Partial migrating from RIP V1 to RIP V2, some
  business or company organizational reason to use
  both versions, etc.)
• 2. Regardless of the reasons, to support both
  versions in the same internetwork, one or more
  routers need to use both versions at the same
  time.

30
RIP Version Migration Speaking Both Versions

• 1. Configure R1 for RIP V1 (by omitting the
  version command) and then configure interface
  S0/0 to send and receive RIP V2 updates
• 2. Configure R1 for RIP V2 (by including the
  version 2 command) and then configure interface
  S0/1 to send and receive RIP V1 updates

31
Configuring RIP Version 2 on an Interface

• R1 enables RIP V2 on interface S0/0 by using the
  ip rip send version 2 and ip rip receive version
  2 interface subcommands.
• So, R1 sends and receives only RIP V2 updates on
  the right of Figure and defaults to sending and
  receiving RIP V1 updates on the left.

32
Design Options Impacted by the RIP V2

• The use of RIP V2 instead of RIP V1 allows the
  use of two powerful network
• design options.
• 1. V2 allows for the use of VLSM. VLSM gives the
  engineer much more
• flexibility when choosing which subnets to use
  and how many hosts to put into
• each subnet.
• 2. RIP V2 also allows a design choice called a
  discontiguous network. A
• discontiguous network occurs when at least one
  pair of subnets of the same
• classful network are separated by subnets of a
  different classful network.
• RIP V1 does not support discontiguous networks
  RIP V2 supports them if all
• the routers have been configured with the RIP no
  auto-summary
• subcommand.

33
Discontiguous Network 172.30.0.0
34
Other RIP Configuration Options

• RIP has several optional configuration settings
  as well
• Adjust timers, such as the holddown and update
  timers
• Enable or disable split horizon per interface
• Explicitly configure RIP neighbors to support
  certain types of WAN connections
• Disable the sending of RIP updates on an
  interface (using the passive-interface command),
  while still receiving RIP updates
• Filter the contents of RIP updates

35
RIP Timers

• RIP uses several timers
• 1. Update timer
• 2. Holddown timer
• RIP uses the concept of an invalid timer and a
  flush timer. (The flush
• timer determines when a router removes a route
  from the routing table
• after the route has been poisoned.)
• All of these timers can be reset with the
  following command, which is
• configured as a subcommand under router rip
• timers basic update invalid holddown flush
• You might consider lowering the holddown timer to
  speed convergence.

36
Disabling Split Horizon

• Split horizon helps prevent loops by avoiding the
  counting-to-infinity
• problem.
• Cisco IOS enables split horizon on all interfaces
  
• (except serial interfaces that are configured
  with some of Frame Relay
• options). However, you can disable split horizon,
  per interface using
• interface subcommand
• no ip split-horizon
• For example, to disable split horizon on
  interface S0/0, the engineer
• would enter configuration mode, type the
  interface S0/0 command,
• and then use the no ip split-horizon command.

37
Configuring Neighbors

• RIP V1 sends its update messages to IP broadcast
  address 255.255.255.255.
• RIP V2 Improves the update process by sending its
  update messages to the
• 224.0.0.9 multicast IP address.
• By using multicasts, only RIP-speaking routers
  should process RIP updates,
• reducing the overhead on the other hosts on a
  LAN.
• However, some WAN data links may not support the
  sending of data-link
• broadcasts or multicasts.
• In those cases, RIP must send its updates using
  IP and datalink unicast
• addresses.
• To do so, a RIP router must define the
  neighboring routers unicast IP address
• using the neighbor command under router rip.

38
Enabling the passive-interface Command

• After RIP is configured, it may be useful to then
  stop sending RIP
• updates on the interface.
• To do so, the configuration must still match the
  interface with a
• network command, and then the router must be told
  to stop sending
• updates with the passive-interface interface
  subcommand under
• router rip.
• The passive-interface command tells RIP to stop
  sending RIP updates
• out the listed interface.

39
Route Redistribution
R2, which uses only OSPF, has no need to receive
R1s RIP updates. So, R1 has used the
passive-interface command, meaning that R1 no
longer sends RIP updates out its S0/0 interface
into the OSPF part of the internetwork.
40
Filtering Routes

• Routers can use route filtering to filter the
  routes sent and
• received in RIP updates.
• Route filtering allows an engineer to limit which
  routers
• learn which routes.
• For example, if a particular subnet should be
  protected for security
• reasons, and only certain groups of people should
  be able to
• communicate with the hosts in that subnet, the
  engineer could filter
• routes.

41
Verifying RIP Operations Using show Commands

• The following four show commands provide the most
  
• useful information for examining how RIP is
  working in a
• router
• - show ip protocols
• - show ip route
• - show ip interface brief
• - show ip rip database

42
R1 Sample RIP show ip protocols Command
43
Sample RIP show Commands on R1
44
Troubleshooting RIP Operations Using the debug
Command

• Cisco IOS supports a very important
  troubleshooting command called
• the debug command.
• The debug command has many options, including
  options related to RIP.
• Regardless of what options are added to the debug
  command, this
• command tells the router to do the following
• - Monitor some internal process (for example,
  RIP updates that are sent and received)
• - When something happens related to that
  process, generate log messages
• - Keep generating log messages until someone
  disables the debug using the no debug command

45
R1 Messages Generated by the debug ip rip Command
46
Load Balancing over Multiple Equal-Cost Routes

• When a router discovers multiple equal-cost
  routes to the same subnet,
• using a single routing protocol, the routing
  protocol can add multiple of
• those routes to the routing table.
• All the IGPs on Cisco routers use the following
  (default) rules when
• considering multiple equal-cost routes
• - By default, add up to four equal-cost
  (equal-metric) routes for the same subnet to the
  routing table at the same time.
• - The number of concurrent equal-cost routes can
  be changed by using the maximum-paths number
  subcommand, to a value between 1 and 6.

47
Load Balancing over Multiple Equal-Cost Routes
(Continued)

• When the IP routing table lists multiple routes
  to the same destination,
• the IP routing process then needs to choose how
  to load-balance the
• traffic over the multiple routes.
• The following two options based on the internal
  routing process used
• by the router
• - Process switching The slowest and
  highest-overhead option for how IOS forwards
  packets. However, with process switching, load
  balancing occurs per packet, with each successive
  packet going to the destination subnet using a
  different route.
• - Fast switching The next fastest option, with
  less overhead, for how IOS forwards packets.
  However, when using fast switching, the router
  balances traffic per destination IP address.

48
Equal-Cost Load Balancing
R1 Messages Generated by the show ip route
Command
49
Choosing Routes Based on Administrative Distance

• In some cases, one router may need to use
  multiple routing protocols.
• Because each routing protocol uses a different
  metric, a router cannot
• use the metric to determine which route is the
  best route.
• Routers determine the best route in these cases
  by choosing the route
• with the lowest administrative distance.
• The administrative distance is a number assigned
  to all the possible
• sources of routing information routing
  protocols and static routes
• included.

50
Default Administrative Distances in Cisco IOS
51
Floating Static Routes

• A floating static route is a static route that
  the engineer wants to be
• used some of the time. The term floating comes
  from the idea that the
• static route leaves the routing table under some
  conditions and comes
• back into the routing table under other
  conditions.
• Floating static routes can be very useful for
  dial backup, using the
• following logic
• - When a permanent WAN connection is up, the
  router should ignore the static route and instead
  use the routes learned by the routing protocol.
  These routes will forward packets out the
  permanent WAN connection.
• - When the permanent WAN connection is down, use
  the statically defined route that sends traffic
  over the dial backup link.

52
Advertising Default Routes with RIP

• In some cases, it makes sense to distribute a
  default route throughout an internetwork.
• 1. All routers in the enterprise internetwork
  learn about all subnets of Class B network
  130.1.0.0 via RIP.
• 2. Router R-core defines a static default route
  pointing to the Internet.
• 3. Router R-core advertises a default route to
  the rest of the routers in the enterprise.

53
Classless and Classful Routing Protocols

• The term classless routing protocol refers to a
  set of routing
• protocols that provide a particular set of
  functions.
• Classless routing protocols perform the following
  functions
• - Send subnet mask information in routing
  updates
• - Support variable-length subnet mask (VLSM)
  because of the inclusion of the mask in routing
  updates
• - Support designs that include discontiguous
  networks
• A classful routing protocol, by definition, does
  not send mask
• information. As a result, it does not support
  VLSM, nor does it support
• discontiguous networks.

54
Classless and Classful Routing

• The terms classful routing and classless routing
  refer to how each
• router uses its default route, assuming the
  router has a default route.
• - Classless routingIf a packets destination IP
  address does not match a more specific route in
  the IP routing table, forward the packet based on
  the default route.
• - Classful routingIf a packets destination IP
  address does not match a more specific route in
  the IP routing table, forward the packet based on
  the default route, but only if the routing table
  does not contain any subnets of that packets
  classful IP network.

55
Classless and Classful Addressing

• The terms classless addressing and classful
  addressing refer to two
• methods of analyzing the structure of IP
  addresses.
• Classful addressing means that, when analyzing IP
  addresses, the
• addresses are considered to have a one-, two-, or
  three-octet network
• part, with the remainder of the addresses being
  the host part.
• Classless addressing ignores Class A, B, and C
  rules, treating each
• IP address as having only two parts a subnet
  part and a host part. In a
• classless address, the subnet part (also called
  the prefix) contains what
• would have been the combined network and subnet
  parts with classful
• addressing.

56
Classless and Classful Addressing Compared
57
Summary

Distance vector routing protocols use a wide
variety of loop-avoidance features.
58
Summary (Continued)
59
Summary (Continued)

• Distance vector algorithms call for each router
  to send its entire
• routing table, on a periodic basis, to each of
  its adjacent neighbors. The
• routing tables include information about each
  network or subnet, along
• with the metric associated with each network or
  subnet.
• RIP Version 2 added many features to RIP Version
  1. These
• enhancements include an authentication mechanism,
  support of VLSM,
• and support of discontiguous networks.
• The two most common commands used to verify that
  RIP is properly
• configured are the show ip route and show ip
  protocols commands.

60
Summary (End)

• The passive-interface command prevents routers
  from sending
• routing updates through a router interface.
• RIP supports the function of adding multiple
  equal-cost routes to the
• same subnetup to four routes by default, and up
  to six possible.
• Classless routing, as enabled with the ip
  classless global
• configuration command, means that a router always
  uses its default
• route (assuming one exists) if a packets
  destination address does
• not match another route.
• With classful routing enabled (with the no ip
  classless global
• command), the router may discard some packets
  even if a default route
• exists.