SlideShare une entreprise Scribd logo
1  sur  53
Télécharger pour lire hors ligne
Routing Architecture




                                                 Module 2
                                                 Routing Fundamentals

                                                  Basic Problems
                                                  Principles, Classification
                                                  Operation




Author: Rolf Augstein
        raugstein@rolfaugstein.com
        January 2006


         Feel free to use this publication for private, non-commercial purposes.




   Objectives

         1.     Basic understanding of routing graphs
         2.     Describe the process of routing through a given network
         3.     Identify problems with Distance-Vector and Link-State protocols
         4.     Understand the solution for different routing problems
         5.     Outline different routing classifications
         6.     Describe the process of route summarization
         7.     Understand the relationship IP addressing scheme - routing functionality




Rolf Augstein                                     © 2006 All rights reserved         Page 1
Key terms:

        •       Aggregate Route
        •       Classless Inter-Domain Routing (CIDR)
        •       Classless Routing
        •       Convergence
        •       Count-to-Infinity
        •       Distance Vector (DV)
        •       Exterior Routing Protocol EGP)
        •       Flapping Route
        •       Floating Static
        •       Fixed Length Subnet Mask (FLSM)
        •       Interior Routing Protocol (IGP)
        •       Link State (LS)
        •       Metric
        •       Poison Reverse
        •       Preference Value
        •       Prefix Routing
        •       Route Summarization
        •       Routing Hierarchy
        •       Routing Loops
        •       Smart Router
        •       Split Horizon
        •       Variable Length Subnet Mask (VLSM)




Rolf Augstein                              © 2006 All rights reserved   Page 2
Routing Principles
       Routing in general is a method of finding the best way through a given network
       of roads or rail-tracks, for example. The term “best way” depends on individual
       parameters. It could mean the fastest, cheapest, or most comfortable one.
       Mathematical algorithms like “Dijkstra”, are used to find out the “best” way
       through a given network. The discipline dealing with this kind of problems is
       called the graph theory.



         Graph

       Graphs are used to show all possible ways from a source to a destination. Not
       all combinations of ways are possible in the typical graph below.

       Example:

            •    It is not possible to go directly from node C to node B
            •    You can go from node B to node F, but not same way back




                Theory of graphs

                                                                      8                   H
                                                 D
                                                                  3               3
                                         3                                E
                                                     13
                             C
                                                     11
                                                          F       1
                                                                                      7
                                                                              4

                         6       5
                                                     3
                                                              2


                                     2       B                        G
                     A

                                                      - Where are the possible paths ?
                         From A to H:
                                                      - What´s the cost for each path ?
                    Find the best way
                                                      - What´s the best path ?




Rolf Augstein                                    © 2006 All rights reserved                   Page 3
Further, there are different metric values through certain paths between two
       nodes. The metric value from node A to node C is 6. The opposite direction,
       node C to node A has a metric value of 5 only.




       Different elements are used to draw relations between certain nodes.




                Elements of Graphs

                                                                    Examples:
                          3


                                                                Serial Links, Shared Medium, etc.
                - both directions, equal cost



                          5


                                                                Special Links (Satellite)
                - one direction



                              10


                              3                                 ADSL

                - both directions, unequal cost




Rolf Augstein                                     © 2006 All rights reserved                    Page 4
Important Terms

       A graph consists of vertices (nodes) and edges. Two vertices are adjacent, if
       they are connected by an edge.




       Example: This is a graph with 6 vertices (nodes) and 7 edges.


       A graph is called a complete graph, if each edge is connected to each of the
       others in the graph. Below are the first 5 complete graphs.




       In data networking, this kind of graph is often called a “fully meshed network”.
       The number of edges in a complete graph is increasing dramatically with each
       new node. The formula to calculate the number of edges (possible ways) in a
       fully meshed network is:

                                           n * ( n-1)

                                                2



       One more important term is directed graph. A digraph (directed graph) is a
       graph where edges are directed. This means that there are only certain possible
       ways through the graph.




       The arrows mark the direction from which the graph is determined. In this
       example, we have a complete graph but there is no direct path from C to B.




Rolf Augstein                             © 2006 All rights reserved                Page 5
Basic Routing Topologies


       Because of the size of modern data networks, it is not possible to connect each
       node with all other nodes. So, fully meshed networks are normally not subject of
       a network design.




                Figures




                                                                      Partially meshed


                  Fully meshed




                                    Partially meshed, Hub-and-Spoke




       Fully meshed networks can be found in parts of Wide Area Networks like ATM,
       Frame Relay or X.25. On the other hand a meshed network is more reliable
       because of the redundancy. But the routing becomes very complex.

       In IP data networks each node is represented by an IP Router or a Switch with
       Layer 3 capabilities. From this point of view, the graphs draw the network
       topology from the IP layer.

       In most cases the design is based on partially meshed networks. This means
       not all nodes are connected to each other. The design is more based on
       geographic issues or available bandwidth etc.

       The Hub-and-Spoke architecture is often used where smaller locations like
       SOHO or ROBO are connected to a centralized node.




Rolf Augstein                            © 2006 All rights reserved                      Page 6
Metric



                Criteria for finding the best way




                  ?                                                     ?

                                       Path length              Reliability
                 Metric                Cost factor               ....

                                       Bandwidth




       The value metric is used in all routing procedures or protocols. In most cases
       the metric represents nothing more than an abstract value. Depending on the
       routing procedures, the metric has different meanings. Sometimes the metric
       counts the number of hops between two nodes. In other cases the metric is
       calculated out of the available bandwidths on the path, the delay, the MTU, the
       load, or the communication cost.



       Note

       The smaller the calculated metric, the better the way is. This is true for all
       dynamic and static routing procedures.



       Different routing protocols use different metric calculations. For this reason
       there is no compatibility between the metric values of dynamic routing protocols.
       To overcome this problem, it is possible to use route redistribution. This is
       covered later in this module.




Rolf Augstein                              © 2006 All rights reserved                   Page 7
Each routing protocol uses a default method to calculate the metric between the
       nodes. For the network administrator it is possible to influence and manipulate
       the metric calculation and the way these information are passed between
       neighbours.


       It is possible to alter the entire routing behaviour in a given network disregard of
       real physical structure and cabling.



       Note

       Therefore the administrator must clearly understand all aspects of the routing
       protocol and it’s behaviour. Do not change metric values in a complex network
       structure just to “find out”.




       You can the force the entire data flow through a network to take different paths
       for special settings.




       Example: Asymmetric Routing

       Packets to the destination use a different path than the packets back from the
       destination. This is called asymmetric routing. With manipulation of the routing
       metrics, a router becomes an altered directed graph, a new logical topology of
       the data network.




Rolf Augstein                              © 2006 All rights reserved                 Page 8
Routing Classification

       There are some different ways to make a routing classification. Three are
       covered in the following.



         Static Routing vs. dynamic Routing


                     Static vs. Dynamic




                                 Hmm. Which
                              networks do I need
                                  to reach ?
                                                                                                     Tell me, which
                                                                                                      networks are
                                                                                                        available




                         Mr. Administrator

                         Defined by Administrator                                     Learned from Network




           Developed by Media-Learning.com © 2005 All rights reserved                                                 2-8




       Static Routing

       With static routing all destination networks and useable paths must be defined
       in the router. These definitions are a big administrative challenge.

       All routers at the remote side must have a route back the originating network.
       The router does not learn which data packets of a session were routed earlier.
       The return packet within a session must be routed back – therefore a back route
       has to be defined as well.

       Nevertheless, static routing still plays a role even in big networks.




Rolf Augstein                                                           © 2006 All rights reserved                     Page 9
Table: Static Routing



                            Advantage                                      Disadvantage

                No routing updates, less traffic             No adaptation when links change
                Compatible with each router-system           Complexity in bigger networks
                No flapping routes




       Dynamic Routing

       Dynamic routing or adaptive routing uses protocol updates to propagate all
       known networks to all adjacent nodes. All possible paths through the data
       network are explored and learned. So the routers can take advantage of
       redundant links and react automatically whenever a link between two nodes is
       lost.

       Even the back routes are learned through this dynamic mechanism. There are
       various dynamic routing protocols with different level of complexity.


       Table: Dynamic Routing



                            Advantage                                      Disadvantage

                All paths are propagated dynamically         Routing Updates cause traffic
                Adaptation when links change                 Convergence, “ugly” route effects*
                                                             More administrative knowledge



       An administrator must have enough knowledge regarding the update behavior
       between the routing nodes. They are quite different and come with tricky
       problems and solutions. Examples are “Count-to-Infinity”, Split Horizon etc.*




       * Problems and effects with Routing Protocols are covered later in detail.




Rolf Augstein                                 © 2006 All rights reserved                  Page 10
Destination Routing vs. Source Routing



                Routing: Destination vs. Source



                  Destination Routing                                        Source Routing



                Routing decision based on                             Routing decision based on
                IP Network to go                                      the Source of the IP packet



                Examples: RIP, OSPF                                    Examples: Policy Routing




                    Arriving IP Packet   Destination IP   Source IP   Data




       Whenever a data packet arrives at the router, the destination IP address is
       checked against the routing table. If the destination network address is not
       defined in the routing table, the packet will be dropped.

       When working with static routing or dynamic routing protocols, this is the default
       procedure for the IP router in most cases. Routing protocols like RIP, OSPF etc.
       are based on destination routing.


       It is also possible to use the source part of an IP data packet to make a routing
       decision. For this to make work, an administrator must define special route
       maps.


       Example:

       A route map defines to forward all data packets from the network 10.12.5.0/24
       to the Ethernet interface 3, and all data packets from 10.12.6.0/24 to the next
       hop gateway 10.10.45.1.




Rolf Augstein                                        © 2006 All rights reserved                     Page 11
The example shows no use of any destination IP addresses to make the routing
       decision. Routing decisions are not longer based on best paths with low metric.
       Routing becomes a matter of local policies.



       Note

       This is also called policy based routing. An administrative policy rules how
       routing decisions have to be made.




       When using this kind of routing, all rules for routing data traffic are defined
       statically in route maps. When the network becomes bigger, it can be very
       difficult to avoid “loosing routes somewhere in the network”.


       It is possible to combine destination routing with source routing within a routing
       node. Source routing is often used in conjunction with Quality of Service (QoS).




Rolf Augstein                              © 2006 All rights reserved                    Page 12
Interior Routing vs. Exterior Routing



                Interior vs. Exterior Routing Protocols




                                  IGP                                   AS 56



                                        EGP




                AS 53                       IGP




       In larger networks it is necessary to use special routing protocols to handle the
       huge amount of routing information.

       Interior Gateway Protocols

       These protocols are used within an administrative area called Autonomous
       System (AS). Within an AS an administrator can decide with routing policy to
       use. Two or more Autonomous Systems can be linked together with the help of
       border routers.

       Typical routing protocols are:

            RIP Version 1/ 2        Routing Information Protocol
            OSPF                    Open Shortest Path First
            IS – IS                 Intermediate State – Intermediate State
            Cisco IGRP              Interior Gateway Routing Protocol
            Cisco EIGRP             Enhanced EIGRP




Rolf Augstein                              © 2006 All rights reserved              Page 13
Note

       Autonomous Systems are identified by a 16 Bit number. This number is
       administrated from the Internet Assigned Numbers Authority (IANA).

       Two Internet RFCs discuss autonomous systems: RFC 1930 (Guidelines for
       creation, selection, and registration of an Autonomous System, March 1996)
       and RFC 0975 (Autonomous confederations, February 1986)

       According to RFC 1930 , "Without exception, an AS must have only one routing
       policy. Here routing policy refers to how the rest of the Internet makes routing
       decisions based on information from your AS."




       Exterior Gateway Protocols


       With an Exterior Gateway Protocol capsulated routing information within one
       Autonomous System is send to a second AS. The EGP connects Autonomous
       Systems by delivering dynamic procedures to propagate routing changes in a
       controlled manner.


       Typical routing protocols are:

            EGP                    Exterior Gateway Protocol (Old, barely used)
            BGP                    Border Gateway Protocol



       BGP design and configuration can be very complex. It is mostly used in some
       internet areas where carriers and internet providers are working together.




Rolf Augstein                             © 2006 All rights reserved              Page 14
Routing Operation

         Finding the Way



                Routing Tables



                                              1                  Net    Gateway
                                    1a                            1      Direct
                                                                  2      Direct
                          a                                       3       2b
                                                                  4       2b

                                    2a
                                                        2
                                                         2b                          Net   Gateway
                                                                                      1      2a
                                                  b                                   2     Direct
                                                                                      3     Direct
                                                                                      4      3c
                                                         3b             3
                              Net   Gateway                                 3c
                               1      3b
                               2      3b
                               3     Direct
                                                                                 c
                               4     Direct
                                                                       4c            4




       The basic idea behind routing protocols is, to send local routing information to
       adjacent routing nodes.


       All connected interfaces with a configured IP address, cause an entry in the
       local routing table. The routing table consists of information to reachable
       destination networks. Local networks are marked as “direct connected” or
       “local”.


       With routing update packets send in a given time interval, neighbor routers
       using the same routing protocol learn possible ways to IP networks. In the next
       step, all learned routes from adjacent routing nodes are sent again in the next
       update cycle.




Rolf Augstein                                     © 2006 All rights reserved                         Page 15
If a routing node learns routes via OSPF routing, these routes are not updated
       by a different routing protocol like RIP. To make these protocols to interact,
       route redistribution is necessary.




                Next Hop


                                                                1
                                                          1a

                                                     a

                I can reach network 3 and                 2a
                                                                         2
                4 through my “Next Hop”,                                 2b
                router 2b
                                                                    b

                                                                         3b
                                                                               3
                                                                                   3c

                  Next Hop:                                                             c

                                                                              4c
                                                                                            4
                     Interface IP Address of the
                     directly connected neighbor router




       All reachable IP destination networks are learned in the routing table. But the
       router has only a limited number of information sources.



       Example:


       Router “a” has only one information source, which is the adjacent router “b”.
       There is no information, telling router “a” that there is a third router “c”. But
       router “a” can reach the network “4” through router “b” as well.

       This information source is called the “next-hop gateway”. So after some time, a
       router learns all reachable networks, but is not aware of all other routers in the
       network. This is sometimes referred to as “routers have a flat view of the
       network”.

       Routers must have a valid route to the next-hop gateway. So always use the
       directly connected interface of the next-hop gateway as IP address.




Rolf Augstein                               © 2006 All rights reserved                          Page 16
Bellman Ford Algorithm


       The Bellman Ford algorithm is used to find the shortest way in a graph and is
       the basis of distance-vector routing protocols.




                Distance Vector Routing Protocol




                            RT             RT                RT              RT




                                 Interval n+2      Interval n+1     Interval n



                                           Broadcast Load

                                                           Convergence Problem

                                      Metric Restrictions




       Distance Vector (DV) Routing

       The principle of DV routing is to send routing updates in a defined interval
       through all interfaces. These update packets use broadcast addresses, and
       contain information about all the reachable networks.

       The vector consists of the source address of the sending router. By this address
       receiving routers learn the address of the next-hop gateway over which the
       propagated networks can be reached.

       The distance describes the metric. In most cases this is simply the number of
       hops to a destination network. This number is restricted to a maximum of 15
       hops.

       A typical DV protocol is the Routing Information Protocol (RIP). It is widely used
       and implemented in all UNIX and Windows Servers.




Rolf Augstein                                   © 2006 All rights reserved         Page 17
By depending on a fixed time interval to send the routing updates to all
       neighbors, routing information need a certain amount of time to travel through
       the network. This effect is called “convergence”.

       The use of broadcast addresses causes in the WAN part of large network some
       problems.

       The advantage of DV routing is the simple implementation and the easy way to
       use it in networks.




         Link State Algorithm


       Link-State algorithms are the solution for modern routing protocols. But they
       operate in a totally different way than the DV protocols.




                Link-State Routing Protocol




                                 LSA                                                      CPU
                  Topology
                  Database                                                             Memory




                                                                    RT
                                                                  ..................
                                                                  .................
                                                                  .................
                                                                  .................
                                                                  ..........
                                                                  ...............

                        SPF Algorithm


                                           Shortest Path First Tree




Rolf Augstein                             © 2006 All rights reserved                            Page 18
Link State (LS) Routing

       The basic concept of link-state routing is that every node receives a connectivity
       map of the network, in the form of a graph showing which nodes are connected
       to which other nodes. Each node independently calculates the best next hop
       from it for every possible destination in the network


       Each router builds a relationship with all other routers using a link-state
       protocol. Different roles like designated router, area router, border router etc.
       are assigned. Each node periodically makes up a short message, the link-state
       advertisement, (LSA). The LSA´s are used to identify other nodes which are
       directly connected and keep track of changes in routing.


       All information concerning other routing nodes and reachable networks are
       stored in the topology database. Compared to DV routing, a LS router holds
       more information about the entire network and does not have a flat view only.
       To find the best way through all the reachable destination networks, LS routing
       uses the algorithm SPF.


       Shortest Path First (SPF)


       A routing node uses the stored graph to calculate all paths to each other routing
       nodes. The paths with the best metric values are used to forward IP data
       packets. The result is a spanned tree with best paths to all destination networks
       instead of a flat view compared to DV routing.


       The advantage of LS routing is quick reaction to any changes in the network
       topology.




Rolf Augstein                             © 2006 All rights reserved               Page 19
Process Topology Changes



                Link Up-Down




                                                              router% Line protocol down......
                                                                      or
                                       Keepalive Timer        router% Line protocol up......




                                                          Entries in Routing Table
                                              ..........
                                              C 194.123.123.16 is directly connected, Ethernet0
                                              R        network 123.123.0.0 via Ethernet 0
                                              R        network 34.23.0.0 via Ethernet 0
          All routes associated with
          interface Ethernet 0 are            C 193.141.147.0 is directly connected, BRI0
          not valid any longer                ..........




       How does a routing node realize changes in the network topology?

       Usually, topology changes cause error states on the connected router interface.
       The line protocol goes down or the interface hardware fails.

       To control the functionality of the interfaces, the Operating System generates
       control packets which are sent through the interface.

       If the interface signals a problem, the operation state changes and all
       corresponding routes are effected in the routing table.




Rolf Augstein                                © 2006 All rights reserved                           Page 20
Routing Timer



                  Update
                                                               Time between Updates



                                                            network unreachable ....
                  Invalid                                   network possibly down .....

                                                            Time after the entry is
                                                            marked as “invalid”


                  Flushing
                                                             Route is erased from
                                         network 13.2.3.0
                                                             Routing table




       To avoid flapping interfaces and flapping routes the entire state change process
       uses a delay mechanism.



       Note

       The term “flapping” is often used to describe a failure condition, where i.e. an
       interface changes the state between up and down very often in small time
       intervals. This can cause a lot of problems and effect the entire network routing.




       An invalid timer controls when a route is marked as possibly unreachable or
       down. This timer is set 2 – 3 times higher than the update timer. At least two
       missed updates are necessary to cause a change in routing.


       An additionally flushing timer determines when a routing entry marked as
       possibly down is erased out of the routing table.




Rolf Augstein                              © 2006 All rights reserved                     Page 21
Using multiple Paths



                Load Balancing


                             Route A




                             Route B




                Packets are “balanced” through multiple ways

                                       More Bandwidth
                Advantage:
                                       Higher Availability




       When the routing process has two or more paths with equal metric to a
       destination network, it is possible to send the data packets along these routes.

       The data load is balanced. Some routing protocols can perform unequal cost
       load balancing with up to 5 different routes.




Rolf Augstein                             © 2006 All rights reserved               Page 22
Load Balancing (cont.)



                    Problem:

                                                            Route A            Route B
                  Different Trip Times                      114 ms             262 ms




                    - Per Destination Load Balancing

                    - Per Packet Load Balancing




       One of the main problems when performing load balancing exists in the different
       trip times of particular routes. This can cause problems for data application
       when the packets arrive in a different order then actually sent.

       Special care must be taken. Different techniques are available to solve the
       negative effects.


       Example

       A gateway has 2 two different paths to the headquarter network. The first
       session initiated is sent through the first known path, the second session is sent
       through the second path. The third session must use the first path and so on.
       This called “Per Destination Load Balancing”.

       When using “Per Packet Load Balancing” all packets regardless of the session
       ID are balanced over both paths. In this case the load on the different paths is
       balanced in a optimized manner. But the risk of packet delays with a higher rate
       of retransmissions is more likely.




Rolf Augstein                             © 2006 All rights reserved                 Page 23
Control Packet Lifetime



                Time-to-Live


                                 IP-Version 4 Header




                           TTL 23
                            TTL 23                                     TTL 22
                                                                        TTL 22



                                     Decreasing Time-To-Live Counter
                                     when passing through router




       In the header of the IP packet, the field TTL takes care of data packets not
       travelling in the network for ever. Whenever a routing node forwards an IP
       packet, the TTL counter is decreased by one.

       A packet with TTL set to 0 is discarded by the router.




Rolf Augstein                             © 2006 All rights reserved                  Page 24
Routing Problems

       Each routing protocol has advantages and also disadvantages. There is no
       perfect routing protocol. An administrator must deal with the pros and the cons
       trying to find the best solution for his needs.



         Convergence



                 Convergence Problem

                                                                                     New Route to
                                                                                     194.200.1.0




            194.200.1.0    194.200.1.0   194.200.1.0     194.200.1.0        194.200.1.0



                300 secs   240 secs      180 secs       120 secs       Next Update
                                                                       in 60 secs



                              Worse case scenario




       A major problem with DV routing is the convergence problem. New information
       like changes in routing take quite some time to get to all members of the routing
       process.

       The negative effect is increasing when networks become bigger and the
       changes occur much more often.




Rolf Augstein                                  © 2006 All rights reserved                       Page 25
Count to Infinity



                 Count to Infinity




                                             ?
                Don´t worry ! I have a
                route to 194.200.1.0
                                                                                    No Route to
                                                                                    194.200.1.0




           194.200.1.0         194.200.1.0                  194.200.1.0       194.200.1.0




                                    Worse case scenario




       Slow convergence causes additional problems. Routers update routing
       information to neighbours, even if they are the source of this information.

       This phenomenon is called count-to-infinity, because it leads to a ping-pong
       effect until the maximum value for the metric is reached.



       So how can one overcome this kind of effects?




Rolf Augstein                                    © 2006 All rights reserved                 Page 26
Triggered Updates



                 Solution: Triggered Updates


                                                                   interface down




                                      Network unreachable
                                         Metric <max. Value>


                Neighbor receives Update
                with max. Metric
                                                        Overcome
                                                        Convergence Problem !
                Any other Changes are
                transmitted immediately




       The flow of negative information must be accelerated. Whenever a change in
       routing occurs, these changes are transmitted immediately to all adjacencies. If
       an entire network is unreachable, the update packets contain the metric value
       set to the maximum.



       Poison Reverse

       This technical term is used to indicate, that a packet with a higher metric or the
       maximum metric is set and sent along the reverse path trough the network to
       overcome problems like routing loops or count-to-infinity. Poison reverse is a
       triggered update to speed up the convergence of the routing protocol.




Rolf Augstein                              © 2006 All rights reserved               Page 27
Interface Hold-down


                                           Update from neighbor
                 Network
                           154.34
                                  .23.0 d
                                          own

                                                                                Route Table
                                              er
                                           Tim                        Flush network 154.34.23.0 via E0




                >entering hold-down for network 154.34.23.0


           • Accept no further Information for Network 154.34.23.0
             for a certain amount of time

           • Avoidance of Routing-Loops




       A router should not rely on information arriving on an interface that was sent out
       earlier over that interface. When a route is flushed out of the routing table, new
       update packets for a particular route from any neighbour are not accepted for
       some time.

       A router should realize which routes were propagated through the interfaces
       and should not accept some routes backward. Again these kinds of problems
       occur mainly on DV routing protocols on networks with high convergence.




Rolf Augstein                                      © 2006 All rights reserved                       Page 28
Loops



                Routing Loops


                 Mrs. Easy                                                                          Company Intranet with
                                                           198.210.25.0
                                                                                                    different Administrators

                                                                          De
                                                                             fau
                                                     …                          lt R
                                                                                    ou
                                                  ute to

                                                                                      te
                                                                                           to
                                                                                                …
                                                     o
                                               ult R




                                                                                                          200.200.45.0
                                             Def a




                                195.22.5.0

                                                                   Default Route to …




                                                                                                                         Mr. Brainbox

                   Mr. Theory




       Another problem coming up sometimes is a loop in the routing information table.
       A routing loop can be caused by a lack of communication between different
       routing administrators, for example.


       This is a very tricky problem. It looks ridicules – but it is configured very quickly.


       Another source for routing loops is the way DV-protocols like RIP are working
       as seen in previous chapter. The solution to avoid loops is the Split Horizon.




Rolf Augstein                                                      © 2006 All rights reserved                                           Page 29
Split Horizon


       Split-horizon is a common solution to avoid routing loops. A cause for the route
       loop is that the router propagates routing information learned from a neighbour
       to that neighbour back. The idea of the split-horizon is not to send the routing
       information over the interface that has received this routing information.




                Split Horizon

                                                                   Hub and Spoke

                Is not propagated by RIP


                                                                  Propagated by RIP


                                       Dynamic Routing
                                       with RIP
           Can not access
           network 173.25.0.0 !                                      Network 173.25.0.0




                   Problem: Point-to-Multipoint Interfaces




       The Split Horizon problem comes up in switched wide area networks. In a
       switched network, one physical interface is configured with several instances of
       logical interfaces. The logical interfaces deal with the different IP networks. The
       routing process deals with the physical interface. So information learned from
       the way in on this physical interface is not sent out over the same physical
       interface. This is to avoid routing loops.

       So something that was designed to solve a problem now causes another
       problem.

       Administrators must be aware of the split horizon effect in point-to-multipoint
       interfaces to avoid routing misconfiguration.




Rolf Augstein                                  © 2006 All rights reserved                 Page 30
Routing Interoperability

       Many administrators use more than one routing protocol in their network to
       manage various needs. This chapter covers how different routing protocols can
       configured to interact with each other.



         The Routing Order



                The Routing Preference

                                                    OSPF

                                                 Static

                 Priority ?
                                                  RIP
                OSPF          2
                Static        1
                RIP           3



                                  Choice:      Which routing method
                                                   should be used ?




       Different routing protocols can be configured and activated in parallel on a
       router. But there is no interaction between each other. This means RIP gets all
       routing information for the network and a second routing protocol like OSPF
       calculates the best path through the some network as well.


       Question: So what routing paths are preferred by a data packet ?


       Each routing protocol including static routing methods do have an assigned
       priority value by default. This value is called the preference.




Rolf Augstein                            © 2006 All rights reserved              Page 31
Note:

       Cisco uses the same mechanism for routing interaction. This priority value is
       called Administrative Distance.




                Working with Preference

                                                                                100.0.0.0




                        Entries in Routing-Table

                  Network                 By       Metric   Preference

                  Route to 100.0.0.0      OSPF       3          5

                  Route to 100.0.0.0      Static     1          8

                  Route to 100.0.0.0      RIP        3         10




                      Route with best preference value




       If there are several routes to a destination network, the first value checked is the
       preference value. This means the routing procedure with the highest priority is
       checked first. A lower preference value means more trust for the routing source.
       Again, within a routing procedure like OSPF, RIP, or Static, the metric value is
       used to define the best path.

       For customization purposes the preferences can be manually configured. If an
       administrator wants to trust a RIP derivate route more than an OSPF route, the
       default preference must be changed.


       Different manufacturers have different specifications on the preferences/
       administrative distance of the routing protocols.




Rolf Augstein                                      © 2006 All rights reserved               Page 32
The following table shows the default preferences of the routers of Quidway
       series produced by Huawei. In the table, a value of “0" denotes the direct route,
       and a value of "255" denotes any route from an untrustworthy source.


       Table: Default Preference Values for Quidway Series


                  Routing Protocol            Preference
                       DIRECT                        0
                        OSPF                        10
                       STATIC                       60
                          RIP                      100
                     Internal BGP                  130
                  OSPF AS External                 150
                    External BGP                   170
                     UNKNOWN                       255



       Except the direct route, preferences of all dynamic routing protocols can be
       configured manually according to the users' requirement.




Rolf Augstein                             © 2006 All rights reserved               Page 33
Floating Static



                 Floating Static Route

                                                                                              100.0.0.0


                Use preference values to make
                static routes “interactive”
                                                                              ISDN Link,   Serial Link,
                                                                              Backup       128 KB



                                Entries in Routing-Table

                  Network                  By        Metric   Preference      Via

                  Route to 100.0.0.0       RIP         3         10        Serial Link

                  Route to 100.0.0.0       Static      1         20          ISDN




                 If serial links goes down, ISDN backup is triggered by static route




       With the help of the preference one can make a static route more “dynamic”.
       By default, a static route has higher priority than all other dynamic routing
       procedures. One can change the behaviour, so as long as a dynamic route is
       present in the routing table, these routes are preferred. When for some reason
       the dynamic route disappears, the defined static route takes precedence.

       Floating static routes are often used as part of routing concepts with ISDN
       backup links.




Rolf Augstein                                          © 2006 All rights reserved                    Page 34
Route Redistribution



                Route Redistribution


                Routing with OSPF                        Routing with RIP




                           Metric 117
                                            ?               Metric 2
                           Metric 139                       Metric 5
                                        not compatible




       As mentioned earlier, each routing procedure uses proprietary metric
       calculations. To make them working together and exchange routing information,
       Route Redistribution can be used.

       With Route Redistribution, basically each routing procedure can be transferred
       in each other.

       There are a lot of considerations to make, when using redistribution. This entire
       technique is covered in detail in a later chapter. Administrators should have
       deeper understanding of the single routing procedures before using
       redistribution between them.




Rolf Augstein                             © 2006 All rights reserved               Page 35
Redistribution Policy

                                                      Define rules for redistribution


                                    Convert OSPF Routes to
                                    RIP: Starting Metric 4




                                    Convert RIP Routes to
                                    OSPF: Starting Metric 9




       The basic principle with route redistribution consists in the choice for special
       routing nodes in the network, where redistribution should be established.



       Example:

       A set of definitions rule the way, a RIP route is converted and transferred in an
       OSPF route and vice versa.

       OSPF Metric 230 is converted to RIP Metric 4
       RIP Metric 3 is converted to OSPF Metric 9



       All metric conversions must be set with care, so the entire routing information
       context makes sense. Also, the choice of the position of the router in the
       network redistributing routes is relevant.




Rolf Augstein                              © 2006 All rights reserved                Page 36
Routing Design


       A structured network design is the fundament for implementing a useful routing
       strategy. Without a proper IP addressing, there is no way for scalable and
       stable networks.




                 Routing Hierarchy


                                                                       Building
                                                         Internet      Areas, Domains, AS

                Core




                Edge/ Convergence




                Access




       In larger networks with structured network design, some routers will take special
       control and handling of routing updates. To take advantage of different routing
       mechanisms it is very important to have a well administrated IP address
       scheme.

       Smart Router:

       From the routing perspective, some routers are smarter than others. Because of
       the routing information they hold in their routing table, some routers may have a
       more detailed knowledge about the network.

       This is common technique to control the amount of routing information. Small
       routers in the access zone do not need all information about the entire network.




Rolf Augstein                             © 2006 All rights reserved                   Page 37
A good IP address plan implemented in a well-designed network has the
       following characteristics:



         •      Scalability

                                                  Allows for large
                                                  increases in the number
                                                  of supported sites


         •      Predictability

                                                  Exhibits predictable
                                                  behavior and
                                                  performance


         •      Flexibility

                                                  Minimizes the impact of
                                                  routers, additions,
                                                  changes, or removals




Rolf Augstein                          © 2006 All rights reserved              Page 38
The Prefix



                Prefix


                         Prefix                         Host



                         “Classfull” Routes

                    Class A 10.0.0.0/8                         10.0.0.0 255.0.0.0

                    Class B 129.12.0.0/16                      129.12.0.0 255.255.0.0

                    Class C       201.12.23.0/24               201.12.23.0 255.255.255.0



                         “Classless” Routes

                    Class C       201.12.112.0/21              201.12.112.0 255.255.248.0




       For routing purposes, an IP address without a given subnet mask is “worthless”.
       To make routing decisions the subnet mask must always be considered.
       Instead of using the subnet mask in the dotted decimal format, a more
       convenient format is used.

       The Prefix points out, how many bits within the 32 bits of the IP address are
       used as the network part.



       So a prefix of 20 bits for an IP address like 144.37.99.34 means, you deal with a
       class B network 144.37.0.0 performing 4 bit subnetting.




Rolf Augstein                                      © 2006 All rights reserved               Page 39
Summarize Routes



                Route Summarization



                      Prefix                  Host               Subnetting


                      - Gain more routable networks




                      - Search common network bits for summarization



                      Prefix                  Host               Summarization




       The process of divide a network in smaller sub-networks is done by shifting the
       network bits to the right. (see TCP/IP fundamentals).


       When dealing with large networks, it is important to minimize the amount of
       routing information. Less routing information means less routing update traffic
       and less RAM (memory) needed in the router.

       So the process of summarize many sub-networks to one network is called
       Route Summarization. This is done by shifting the network bits to the left.




Rolf Augstein                             © 2006 All rights reserved                 Page 40
IP Address Management

                                          Route Summarization
                                          Route Aggregation

                                                                   132.17.25.0
                          132.1
                               7.0.0
                                    /16                            132.17.26.0

                Only 1 update necessary                            132.17.27.0
                                                                   132.17.28.0

                                                                   132.17.29.0

                  IP subnetworks are auto-summarized
                  based on Class A, B, C addresses




       By default, most routers perform auto summarization for class A, B, or C
       networks. Instead of propagating up to 254 subnets of the network 132.17.0.0
       (132.17.1.0 to 132.17.254.0) the summarized route 132.17.0.0/16 is used. This
       means an enormous improvement for the amount of routing traffic sent to the
       neighbour router.




       Note:

       Sometimes the term Route Aggregation is used. An aggregate route includes
       different sub-networks by using appropriate subnet masks.




Rolf Augstein                             © 2006 All rights reserved             Page 41
Relevant Bits

                                                                             IP Address: 10.20.35.5/ 16

                     10.20.35.5     00001001     00010100     00100011   00000101



                                     16 Bits Prefix: Marks the relevant bits for all routing decisions




                                            10.20.35.5       00001001    00010100    00100011    00000101
                                                                           Logical   “AND”

                                            255.255.0.0      11111111    11111111    00000000    00000000
                What the Router does !

                                            Network:         00001001    00010100    00000000    00000000


                                                                  Bits to care             Don´t care




       Routing nodes need the IP address and the corresponding subnet mask to
       make routing decisions. Each interface needs this information as part of the
       configuration. With the help of the subnet mask and the logical “AND” operation,
       it is a simple process to read out the network part of the IP address.


       The prefix bits define the network relevant bits within an IP address. This is the
       reason for sending the prefix in each routing update when using routing
       protocols like OSPF or RIP version 2, so different subnet masks can be used
       within a single IP network.

       The older RIP version 1 is not capable of using different subnet masks in one
       class A, B, or C network. RIP updates are not aware of network prefix.




Rolf Augstein                                            © 2006 All rights reserved                         Page 42
Sub-Subnet


                                            Host A                      IP Address: 10.20.35.5/ 16




                10.20.35.5/16   00001001   00010100     00100011   00000101          Subnet 10.20




                10.20.35.5/19   00001001   00010100     0010   0011   00000101       Subnet 10.20.32




                                            Host B                      IP Address: 10.20.35.5/ 19




       Without the subnet mask information, it is not possible to determine the location
       of a given host in the network. VLSM is like using an additional subnet for a
       “main subnet”. A “sub subnet” describes how many subnets are used within a
       defined subnet. Working with VLSM is simple math, but can be complex in real
       live.

       Interesting: Only routing nodes with appropriate routing procedures must be
       “aware” of VLSM. So, not all routing protocols can be used. End systems like
       hosts or servers do not have to deal with VLSM. They just have to be
       configured with a proper IP address and mask.




Rolf Augstein                                         © 2006 All rights reserved                     Page 43
VLSM Routing


                                                                   VLSM

                                                                                    172.16.56.0/24

                                 172.16.13.4/30



                172.16.11.0/24                                         172.16.13.8/30


                                             172.16.0.0/16




                  Optimization with use of various prefix subnets




       Variable Length Subnet Mask is often used, to optimize the address space for a
       given class A, B, or C network. There are lots of small networks with few hosts.
       Using large subnets like 8 bits prefix actually wastes a lot of address space.


       Worse case is a PPP link with the need of two valid IP addresses only. With a 8
       bits subnet, there are 252 wasted IP addresses !




Rolf Augstein                                     © 2006 All rights reserved                    Page 44
Variable-Length Subnet Mask


                     172.16.0.0/24
                      172.16.1.0                                 Aggregate Route
                      172.16.2.0
          254         . . . .                             172.16.14.0/30
          Subnets     172.16.14.0
                      . . . .
                      172.16.254.0                         172.16.14.4
                                                           172.16.14.8
                                                                             62
                                                           . . . .
                                                                             Subnets
                                                           172.16.14.252



                Use one subnet to split into smaller VLSM subnets




       A proven way of using VLSM is, to take a certain subnet out of the group of
       available subnets. Apply the new subnet mask i.e. 30 bits, so 62 new subnets
       are addressable. Each new subnet can address two hosts.

       The benefit is the gain of new smaller routable networks, which can be used to
       address PPP links.



       Note:

       VLSM does not mean an increasing of IP addresses at all. As a matter of fact,
       lots of addresses are lost because of broadcasts and network addresses.




Rolf Augstein                            © 2006 All rights reserved                Page 45
Table: Prefix Calculation


CIDR            Netmask            Hosts /           Class         Typical usage
                                   subnet
/8              255.0.0.0          16777216          A             Largest block allocation
                                                                   made by IANA
/9              255.128.0.0        8388608
/10             255.192.0.0        4194304
/11             255.224.0.0        2097152
/12             255.240.0.0        1048576
/13             255.248.0.0        524288
/14             255.252.0.0        262144
/15             255.254.0.0        131072
/16             255.255.0.0        65536             B
/17             255.255.128.0      32768                           ISP / large business
/18             255.255.192.0      16384                           ISP / large business
/19             255.255.224.0      8192                            ISP / large business
/20             255.255.240.0      4096                            Small ISP / large
                                                                   business
/21             255.255.248.0      2048                            Small ISP / large
                                                                   business
/22             255.255.252.0      1024
/23             255.255.254.0      512
/24             255.255.255.0      256               C             Large LAN
/25             255.255.255.128    128                             Large LAN
/26             255.255.255.192    64                              Small LAN
/27             255.255.255.224    32                              Small LAN
/28             255.255.255.240    16                              Small LAN
/29             255.255.255.248    8
/30             255.255.255.252    4                               "Glue network" (point to
                                                                   point links)
/31             255.255.255.254    2                               "Useless Network",
                                                                   proposed for point to
                                                                   point links (RFC 3021)
/32             255.255.255.255    1                               Host route




Rolf Augstein                             © 2006 All rights reserved                 Page 46
Classless Routing


                                                              Contains block of:

                           Defined Summary Route:
                                                                       200.16.168.0
                           200.16.168.0/21                             200.16.169.0
                                                                       200.16.170.0
                                                                       200.16.171.0
                                                                       200.16.172.0
                                                                       200.16.173.0
                                                                       200.16.174.0
                                                                       200.16.175.0



                  CIDR:
                  Classless Inter-Domain Routing




       The IP address space was divided into three main network classes, where each
       class had a fixed network size. The class, the length of the subnet mask and the
       number of hosts on the network, could always be determined from the most
       significant bits of the IP address. Without any other way of specifying the length
       of a subnet mask, routing protocols necessarily used the class of the IP address
       specified in route advertisements to determine the size of the routing prefixes to
       be set up in the routing tables.

       CIDR uses VLSM to allocate IP addresses to subnets according to individual
       needs. Thus the network/host division can occur at any bit boundary in the
       address. The process can be recursive, with a portion of the address space
       being further divided into even smaller portions, through the use of masks which
       cover more bits.

       Because the normal class distinctions are ignored, the new system is called
       classless routing.




Rolf Augstein                             © 2006 All rights reserved                  Page 47
Prefix aggregation


       Another benefit of CIDR is the possibility of routing prefix aggregation. For
       example, sixteen contiguous /24 networks could now be aggregated together,
       and advertised to the outside world as a single /20 route (if the first 20 bits of
       their network addresses match). Two contiguous /20s could then be aggregated
       to a /19, and so forth. This allows a significant reduction in the number of routes
       that had to be advertised over the Internet, preventing 'routing table explosion'
       from overwhelming routers, and stopping the Internet from expanding further.



       When dealing with aggregate routes within the internet the term “Supernet” is
       used sometimes.




       These kinds of routing mechanisms are part of BGP routing. The Border
       Gateway Protocol is discussed more detailed in a later module.



       CIDR is described in:
       RFC 1519 (http://www.ietf.org/rfc/rfc1519.txt)
       Classless Inter-Domain Routing (CIDR): an Address Assignment and
       Aggregation Strategy.

       RFC 1518 (http://www.ietf.org/rfc/rfc1518.txt)
       Architecture for IP Address Allocation with CIDR




Rolf Augstein                             © 2006 All rights reserved                Page 48
Discontinuous Use of Subnets



                                  Routing with RIP, Auto-summarization


            155.10.34.0/24
                                              198.23.24.0/24                    155.10.35.0/24




                               ?
                      Oh fine: I have two routes
                      by RIP to 155.10.0.0




       Another interesting effect comes up with the discontinuous use of a class A, B,
       or C network, which is important to understand for routing administrators.

       Because routers perform auto-summarization on IP network address borders,
       the above situation arises for a router between two networks using
       discontinuous IP address spaces. From the routing perspective, there are two
       paths to the network 155.10.0.0 – with fatal consequences !

       It is not recommended to split IP networks and use them on different
       discontinuous locations.

       To solve the above problem, auto-summary must be disabled on both routers.
       But then, too many routes are propagated through the network cloud, which
       could lead to additional problems.




Rolf Augstein                                      © 2006 All rights reserved                    Page 49
Prefix Matching




                                                         Priority
                                           192.16.3.33 / 32 Host
                                           192.16.3.32 / 27 Subnet
                                           192.16.3.0 / 24 Net
                                           192.16.0.0 / 16 Block Network
                                           0.0.0.0     / 0 Default




                           Rule: “best prefix matches”




       When using subnetting and VLSM in a network, the routing table has various
       entries for a network with different prefix lengths.

       Longest prefix match or best prefix match refers to an algorithm used to decide
       for the best routing entry.

       Because each entry of a routing table may specify a range of addresses, one
       destination address may match more than another routing table entry. The most
       specific table entry, this means the one with smallest host address range, is
       called the longest prefix match.




Rolf Augstein                            © 2006 All rights reserved              Page 50
Module Review


   Summary


       Static routing is still as important as adaptive routing protocols.


       Adaptive routing protocols are divided in Distance-Vector and Link-State
       protocols.


       Routing decisions are based on preferences and metric calculations.


       Network administrators must be aware of different routing problems like Split
       Horizon, Convergence, Loops, or other effects depending on the routing
       protocol.


       Different routing protocols can interact with the help of routing redistribution.


       Network design and appropriate IP addressing schemes are important for fast
       and stable routing.


       The ability for route summarization and aggregation is the key for adaptive
       routing in larger networks.




Rolf Augstein                              © 2006 All rights reserved                 Page 51
Review Question


            1. Outline the difference between metric and preference?




            2. What are common problems of D-V routing protocols?




            3. Build a small table and outline the advantages and disadvantages of
               L-S routing protocols.




            4. What is the meaning of an adjacent router?




Rolf Augstein                            © 2006 All rights reserved             Page 52
5. What is meant by a “Floating Static Route” ?




            6. Describe the rule “best prefix matches” and the relevancy to routing
               protocols.




            7. What kind of topology is Hub-and-Spoke?




            8. Describe the problems arising with a slow convergence.




            9. What is the preference for a direct connected network ? Why ?




            10. What is the meaning of asymmetric routing ?




Rolf Augstein                             © 2006 All rights reserved             Page 53

Contenu connexe

Similaire à Routing primer

Node Path Visualizer Using Shortest Path Algorithms
Node Path Visualizer Using Shortest Path AlgorithmsNode Path Visualizer Using Shortest Path Algorithms
Node Path Visualizer Using Shortest Path AlgorithmsIRJET Journal
 
Crossing patterns in Nonplanar Road networks
Crossing patterns in Nonplanar Road networksCrossing patterns in Nonplanar Road networks
Crossing patterns in Nonplanar Road networksAjinkya Ghadge
 
IRJET- Survey on Implementation of Graph Theory in Routing Protocols of Wired...
IRJET- Survey on Implementation of Graph Theory in Routing Protocols of Wired...IRJET- Survey on Implementation of Graph Theory in Routing Protocols of Wired...
IRJET- Survey on Implementation of Graph Theory in Routing Protocols of Wired...IRJET Journal
 
A Comparison Of Smart Routings In Mobile Ad Hoc Networks(MANETs)
A Comparison Of Smart Routings In Mobile Ad Hoc  Networks(MANETs) A Comparison Of Smart Routings In Mobile Ad Hoc  Networks(MANETs)
A Comparison Of Smart Routings In Mobile Ad Hoc Networks(MANETs) IJMER
 
Graceful labelings
Graceful labelingsGraceful labelings
Graceful labelingsradhikamathy
 
Mobile ad hoc networks – dangling issues of optimal path strategy
Mobile ad hoc networks – dangling issues of optimal path strategyMobile ad hoc networks – dangling issues of optimal path strategy
Mobile ad hoc networks – dangling issues of optimal path strategyAlexander Decker
 
COMPUTER NETWORKS CHAPTER 3 NETWORK LAYER NOTES CSE 3RD year sem 1
COMPUTER NETWORKS CHAPTER 3 NETWORK LAYER NOTES CSE 3RD year sem 1COMPUTER NETWORKS CHAPTER 3 NETWORK LAYER NOTES CSE 3RD year sem 1
COMPUTER NETWORKS CHAPTER 3 NETWORK LAYER NOTES CSE 3RD year sem 1aishwaryaarrao3
 
Distributed Path Computation Using DIV Algorithm
Distributed Path Computation Using DIV AlgorithmDistributed Path Computation Using DIV Algorithm
Distributed Path Computation Using DIV AlgorithmIOSR Journals
 
Physical organization of parallel platforms
Physical organization of parallel platformsPhysical organization of parallel platforms
Physical organization of parallel platformsSyed Zaid Irshad
 
Performance Analysis of Mesh-based NoC’s on Routing Algorithms
Performance Analysis of Mesh-based NoC’s on Routing Algorithms Performance Analysis of Mesh-based NoC’s on Routing Algorithms
Performance Analysis of Mesh-based NoC’s on Routing Algorithms IJECEIAES
 
Graph Data Structure
Graph Data StructureGraph Data Structure
Graph Data StructureKeno benti
 
Performance comparison of XY,OE and DyAd routing algorithm by Load Variation...
Performance comparison of  XY,OE and DyAd routing algorithm by Load Variation...Performance comparison of  XY,OE and DyAd routing algorithm by Load Variation...
Performance comparison of XY,OE and DyAd routing algorithm by Load Variation...Jayesh Kumar Dalal
 
Effect of node mobility onaomdv protocol in manet
Effect of node mobility onaomdv protocol in manetEffect of node mobility onaomdv protocol in manet
Effect of node mobility onaomdv protocol in manetijwmn
 
Effect of mobility models on the performance of multipath routing protocol in...
Effect of mobility models on the performance of multipath routing protocol in...Effect of mobility models on the performance of multipath routing protocol in...
Effect of mobility models on the performance of multipath routing protocol in...csandit
 
ANGLE ROUTING:A FULLY ADAPTIVE PACKET ROUTING FOR NOC
ANGLE ROUTING:A FULLY ADAPTIVE PACKET ROUTING FOR NOCANGLE ROUTING:A FULLY ADAPTIVE PACKET ROUTING FOR NOC
ANGLE ROUTING:A FULLY ADAPTIVE PACKET ROUTING FOR NOCijcsa
 
Ad hoc routing
Ad hoc routingAd hoc routing
Ad hoc routingits
 

Similaire à Routing primer (20)

Node Path Visualizer Using Shortest Path Algorithms
Node Path Visualizer Using Shortest Path AlgorithmsNode Path Visualizer Using Shortest Path Algorithms
Node Path Visualizer Using Shortest Path Algorithms
 
Crossing patterns in Nonplanar Road networks
Crossing patterns in Nonplanar Road networksCrossing patterns in Nonplanar Road networks
Crossing patterns in Nonplanar Road networks
 
IRJET- Survey on Implementation of Graph Theory in Routing Protocols of Wired...
IRJET- Survey on Implementation of Graph Theory in Routing Protocols of Wired...IRJET- Survey on Implementation of Graph Theory in Routing Protocols of Wired...
IRJET- Survey on Implementation of Graph Theory in Routing Protocols of Wired...
 
A Comparison Of Smart Routings In Mobile Ad Hoc Networks(MANETs)
A Comparison Of Smart Routings In Mobile Ad Hoc  Networks(MANETs) A Comparison Of Smart Routings In Mobile Ad Hoc  Networks(MANETs)
A Comparison Of Smart Routings In Mobile Ad Hoc Networks(MANETs)
 
VLSI routing
VLSI routingVLSI routing
VLSI routing
 
Graceful labelings
Graceful labelingsGraceful labelings
Graceful labelings
 
Mobile ad hoc networks – dangling issues of optimal path strategy
Mobile ad hoc networks – dangling issues of optimal path strategyMobile ad hoc networks – dangling issues of optimal path strategy
Mobile ad hoc networks – dangling issues of optimal path strategy
 
COMPUTER NETWORKS CHAPTER 3 NETWORK LAYER NOTES CSE 3RD year sem 1
COMPUTER NETWORKS CHAPTER 3 NETWORK LAYER NOTES CSE 3RD year sem 1COMPUTER NETWORKS CHAPTER 3 NETWORK LAYER NOTES CSE 3RD year sem 1
COMPUTER NETWORKS CHAPTER 3 NETWORK LAYER NOTES CSE 3RD year sem 1
 
Distributed Path Computation Using DIV Algorithm
Distributed Path Computation Using DIV AlgorithmDistributed Path Computation Using DIV Algorithm
Distributed Path Computation Using DIV Algorithm
 
C0431320
C0431320C0431320
C0431320
 
Network Layer
Network LayerNetwork Layer
Network Layer
 
Physical organization of parallel platforms
Physical organization of parallel platformsPhysical organization of parallel platforms
Physical organization of parallel platforms
 
Performance Analysis of Mesh-based NoC’s on Routing Algorithms
Performance Analysis of Mesh-based NoC’s on Routing Algorithms Performance Analysis of Mesh-based NoC’s on Routing Algorithms
Performance Analysis of Mesh-based NoC’s on Routing Algorithms
 
G0544650
G0544650G0544650
G0544650
 
Graph Data Structure
Graph Data StructureGraph Data Structure
Graph Data Structure
 
Performance comparison of XY,OE and DyAd routing algorithm by Load Variation...
Performance comparison of  XY,OE and DyAd routing algorithm by Load Variation...Performance comparison of  XY,OE and DyAd routing algorithm by Load Variation...
Performance comparison of XY,OE and DyAd routing algorithm by Load Variation...
 
Effect of node mobility onaomdv protocol in manet
Effect of node mobility onaomdv protocol in manetEffect of node mobility onaomdv protocol in manet
Effect of node mobility onaomdv protocol in manet
 
Effect of mobility models on the performance of multipath routing protocol in...
Effect of mobility models on the performance of multipath routing protocol in...Effect of mobility models on the performance of multipath routing protocol in...
Effect of mobility models on the performance of multipath routing protocol in...
 
ANGLE ROUTING:A FULLY ADAPTIVE PACKET ROUTING FOR NOC
ANGLE ROUTING:A FULLY ADAPTIVE PACKET ROUTING FOR NOCANGLE ROUTING:A FULLY ADAPTIVE PACKET ROUTING FOR NOC
ANGLE ROUTING:A FULLY ADAPTIVE PACKET ROUTING FOR NOC
 
Ad hoc routing
Ad hoc routingAd hoc routing
Ad hoc routing
 

Dernier

Decarbonising Buildings: Making a net-zero built environment a reality
Decarbonising Buildings: Making a net-zero built environment a realityDecarbonising Buildings: Making a net-zero built environment a reality
Decarbonising Buildings: Making a net-zero built environment a realityIES VE
 
Testing tools and AI - ideas what to try with some tool examples
Testing tools and AI - ideas what to try with some tool examplesTesting tools and AI - ideas what to try with some tool examples
Testing tools and AI - ideas what to try with some tool examplesKari Kakkonen
 
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptx
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptxPasskey Providers and Enabling Portability: FIDO Paris Seminar.pptx
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptxLoriGlavin3
 
Modern Roaming for Notes and Nomad – Cheaper Faster Better Stronger
Modern Roaming for Notes and Nomad – Cheaper Faster Better StrongerModern Roaming for Notes and Nomad – Cheaper Faster Better Stronger
Modern Roaming for Notes and Nomad – Cheaper Faster Better Strongerpanagenda
 
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...Wes McKinney
 
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024BookNet Canada
 
Enhancing User Experience - Exploring the Latest Features of Tallyman Axis Lo...
Enhancing User Experience - Exploring the Latest Features of Tallyman Axis Lo...Enhancing User Experience - Exploring the Latest Features of Tallyman Axis Lo...
Enhancing User Experience - Exploring the Latest Features of Tallyman Axis Lo...Scott Andery
 
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptxThe Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptxLoriGlavin3
 
Genislab builds better products and faster go-to-market with Lean project man...
Genislab builds better products and faster go-to-market with Lean project man...Genislab builds better products and faster go-to-market with Lean project man...
Genislab builds better products and faster go-to-market with Lean project man...Farhan Tariq
 
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024BookNet Canada
 
Generative Artificial Intelligence: How generative AI works.pdf
Generative Artificial Intelligence: How generative AI works.pdfGenerative Artificial Intelligence: How generative AI works.pdf
Generative Artificial Intelligence: How generative AI works.pdfIngrid Airi González
 
Sample pptx for embedding into website for demo
Sample pptx for embedding into website for demoSample pptx for embedding into website for demo
Sample pptx for embedding into website for demoHarshalMandlekar2
 
TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024Lonnie McRorey
 
Time Series Foundation Models - current state and future directions
Time Series Foundation Models - current state and future directionsTime Series Foundation Models - current state and future directions
Time Series Foundation Models - current state and future directionsNathaniel Shimoni
 
DevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsDevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsSergiu Bodiu
 
What is DBT - The Ultimate Data Build Tool.pdf
What is DBT - The Ultimate Data Build Tool.pdfWhat is DBT - The Ultimate Data Build Tool.pdf
What is DBT - The Ultimate Data Build Tool.pdfMounikaPolabathina
 
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxThe Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxLoriGlavin3
 
Data governance with Unity Catalog Presentation
Data governance with Unity Catalog PresentationData governance with Unity Catalog Presentation
Data governance with Unity Catalog PresentationKnoldus Inc.
 
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyesHow to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyesThousandEyes
 
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...panagenda
 

Dernier (20)

Decarbonising Buildings: Making a net-zero built environment a reality
Decarbonising Buildings: Making a net-zero built environment a realityDecarbonising Buildings: Making a net-zero built environment a reality
Decarbonising Buildings: Making a net-zero built environment a reality
 
Testing tools and AI - ideas what to try with some tool examples
Testing tools and AI - ideas what to try with some tool examplesTesting tools and AI - ideas what to try with some tool examples
Testing tools and AI - ideas what to try with some tool examples
 
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptx
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptxPasskey Providers and Enabling Portability: FIDO Paris Seminar.pptx
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptx
 
Modern Roaming for Notes and Nomad – Cheaper Faster Better Stronger
Modern Roaming for Notes and Nomad – Cheaper Faster Better StrongerModern Roaming for Notes and Nomad – Cheaper Faster Better Stronger
Modern Roaming for Notes and Nomad – Cheaper Faster Better Stronger
 
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
 
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
 
Enhancing User Experience - Exploring the Latest Features of Tallyman Axis Lo...
Enhancing User Experience - Exploring the Latest Features of Tallyman Axis Lo...Enhancing User Experience - Exploring the Latest Features of Tallyman Axis Lo...
Enhancing User Experience - Exploring the Latest Features of Tallyman Axis Lo...
 
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptxThe Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
 
Genislab builds better products and faster go-to-market with Lean project man...
Genislab builds better products and faster go-to-market with Lean project man...Genislab builds better products and faster go-to-market with Lean project man...
Genislab builds better products and faster go-to-market with Lean project man...
 
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
 
Generative Artificial Intelligence: How generative AI works.pdf
Generative Artificial Intelligence: How generative AI works.pdfGenerative Artificial Intelligence: How generative AI works.pdf
Generative Artificial Intelligence: How generative AI works.pdf
 
Sample pptx for embedding into website for demo
Sample pptx for embedding into website for demoSample pptx for embedding into website for demo
Sample pptx for embedding into website for demo
 
TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024
 
Time Series Foundation Models - current state and future directions
Time Series Foundation Models - current state and future directionsTime Series Foundation Models - current state and future directions
Time Series Foundation Models - current state and future directions
 
DevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platformsDevEX - reference for building teams, processes, and platforms
DevEX - reference for building teams, processes, and platforms
 
What is DBT - The Ultimate Data Build Tool.pdf
What is DBT - The Ultimate Data Build Tool.pdfWhat is DBT - The Ultimate Data Build Tool.pdf
What is DBT - The Ultimate Data Build Tool.pdf
 
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxThe Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
 
Data governance with Unity Catalog Presentation
Data governance with Unity Catalog PresentationData governance with Unity Catalog Presentation
Data governance with Unity Catalog Presentation
 
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyesHow to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
 
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
Why device, WIFI, and ISP insights are crucial to supporting remote Microsoft...
 

Routing primer

  • 1. Routing Architecture Module 2 Routing Fundamentals Basic Problems Principles, Classification Operation Author: Rolf Augstein raugstein@rolfaugstein.com January 2006 Feel free to use this publication for private, non-commercial purposes. Objectives 1. Basic understanding of routing graphs 2. Describe the process of routing through a given network 3. Identify problems with Distance-Vector and Link-State protocols 4. Understand the solution for different routing problems 5. Outline different routing classifications 6. Describe the process of route summarization 7. Understand the relationship IP addressing scheme - routing functionality Rolf Augstein © 2006 All rights reserved Page 1
  • 2. Key terms: • Aggregate Route • Classless Inter-Domain Routing (CIDR) • Classless Routing • Convergence • Count-to-Infinity • Distance Vector (DV) • Exterior Routing Protocol EGP) • Flapping Route • Floating Static • Fixed Length Subnet Mask (FLSM) • Interior Routing Protocol (IGP) • Link State (LS) • Metric • Poison Reverse • Preference Value • Prefix Routing • Route Summarization • Routing Hierarchy • Routing Loops • Smart Router • Split Horizon • Variable Length Subnet Mask (VLSM) Rolf Augstein © 2006 All rights reserved Page 2
  • 3. Routing Principles Routing in general is a method of finding the best way through a given network of roads or rail-tracks, for example. The term “best way” depends on individual parameters. It could mean the fastest, cheapest, or most comfortable one. Mathematical algorithms like “Dijkstra”, are used to find out the “best” way through a given network. The discipline dealing with this kind of problems is called the graph theory. Graph Graphs are used to show all possible ways from a source to a destination. Not all combinations of ways are possible in the typical graph below. Example: • It is not possible to go directly from node C to node B • You can go from node B to node F, but not same way back Theory of graphs 8 H D 3 3 3 E 13 C 11 F 1 7 4 6 5 3 2 2 B G A - Where are the possible paths ? From A to H: - What´s the cost for each path ? Find the best way - What´s the best path ? Rolf Augstein © 2006 All rights reserved Page 3
  • 4. Further, there are different metric values through certain paths between two nodes. The metric value from node A to node C is 6. The opposite direction, node C to node A has a metric value of 5 only. Different elements are used to draw relations between certain nodes. Elements of Graphs Examples: 3 Serial Links, Shared Medium, etc. - both directions, equal cost 5 Special Links (Satellite) - one direction 10 3 ADSL - both directions, unequal cost Rolf Augstein © 2006 All rights reserved Page 4
  • 5. Important Terms A graph consists of vertices (nodes) and edges. Two vertices are adjacent, if they are connected by an edge. Example: This is a graph with 6 vertices (nodes) and 7 edges. A graph is called a complete graph, if each edge is connected to each of the others in the graph. Below are the first 5 complete graphs. In data networking, this kind of graph is often called a “fully meshed network”. The number of edges in a complete graph is increasing dramatically with each new node. The formula to calculate the number of edges (possible ways) in a fully meshed network is: n * ( n-1) 2 One more important term is directed graph. A digraph (directed graph) is a graph where edges are directed. This means that there are only certain possible ways through the graph. The arrows mark the direction from which the graph is determined. In this example, we have a complete graph but there is no direct path from C to B. Rolf Augstein © 2006 All rights reserved Page 5
  • 6. Basic Routing Topologies Because of the size of modern data networks, it is not possible to connect each node with all other nodes. So, fully meshed networks are normally not subject of a network design. Figures Partially meshed Fully meshed Partially meshed, Hub-and-Spoke Fully meshed networks can be found in parts of Wide Area Networks like ATM, Frame Relay or X.25. On the other hand a meshed network is more reliable because of the redundancy. But the routing becomes very complex. In IP data networks each node is represented by an IP Router or a Switch with Layer 3 capabilities. From this point of view, the graphs draw the network topology from the IP layer. In most cases the design is based on partially meshed networks. This means not all nodes are connected to each other. The design is more based on geographic issues or available bandwidth etc. The Hub-and-Spoke architecture is often used where smaller locations like SOHO or ROBO are connected to a centralized node. Rolf Augstein © 2006 All rights reserved Page 6
  • 7. Metric Criteria for finding the best way ? ? Path length Reliability Metric Cost factor .... Bandwidth The value metric is used in all routing procedures or protocols. In most cases the metric represents nothing more than an abstract value. Depending on the routing procedures, the metric has different meanings. Sometimes the metric counts the number of hops between two nodes. In other cases the metric is calculated out of the available bandwidths on the path, the delay, the MTU, the load, or the communication cost. Note The smaller the calculated metric, the better the way is. This is true for all dynamic and static routing procedures. Different routing protocols use different metric calculations. For this reason there is no compatibility between the metric values of dynamic routing protocols. To overcome this problem, it is possible to use route redistribution. This is covered later in this module. Rolf Augstein © 2006 All rights reserved Page 7
  • 8. Each routing protocol uses a default method to calculate the metric between the nodes. For the network administrator it is possible to influence and manipulate the metric calculation and the way these information are passed between neighbours. It is possible to alter the entire routing behaviour in a given network disregard of real physical structure and cabling. Note Therefore the administrator must clearly understand all aspects of the routing protocol and it’s behaviour. Do not change metric values in a complex network structure just to “find out”. You can the force the entire data flow through a network to take different paths for special settings. Example: Asymmetric Routing Packets to the destination use a different path than the packets back from the destination. This is called asymmetric routing. With manipulation of the routing metrics, a router becomes an altered directed graph, a new logical topology of the data network. Rolf Augstein © 2006 All rights reserved Page 8
  • 9. Routing Classification There are some different ways to make a routing classification. Three are covered in the following. Static Routing vs. dynamic Routing Static vs. Dynamic Hmm. Which networks do I need to reach ? Tell me, which networks are available Mr. Administrator Defined by Administrator Learned from Network Developed by Media-Learning.com © 2005 All rights reserved 2-8 Static Routing With static routing all destination networks and useable paths must be defined in the router. These definitions are a big administrative challenge. All routers at the remote side must have a route back the originating network. The router does not learn which data packets of a session were routed earlier. The return packet within a session must be routed back – therefore a back route has to be defined as well. Nevertheless, static routing still plays a role even in big networks. Rolf Augstein © 2006 All rights reserved Page 9
  • 10. Table: Static Routing Advantage Disadvantage No routing updates, less traffic No adaptation when links change Compatible with each router-system Complexity in bigger networks No flapping routes Dynamic Routing Dynamic routing or adaptive routing uses protocol updates to propagate all known networks to all adjacent nodes. All possible paths through the data network are explored and learned. So the routers can take advantage of redundant links and react automatically whenever a link between two nodes is lost. Even the back routes are learned through this dynamic mechanism. There are various dynamic routing protocols with different level of complexity. Table: Dynamic Routing Advantage Disadvantage All paths are propagated dynamically Routing Updates cause traffic Adaptation when links change Convergence, “ugly” route effects* More administrative knowledge An administrator must have enough knowledge regarding the update behavior between the routing nodes. They are quite different and come with tricky problems and solutions. Examples are “Count-to-Infinity”, Split Horizon etc.* * Problems and effects with Routing Protocols are covered later in detail. Rolf Augstein © 2006 All rights reserved Page 10
  • 11. Destination Routing vs. Source Routing Routing: Destination vs. Source Destination Routing Source Routing Routing decision based on Routing decision based on IP Network to go the Source of the IP packet Examples: RIP, OSPF Examples: Policy Routing Arriving IP Packet Destination IP Source IP Data Whenever a data packet arrives at the router, the destination IP address is checked against the routing table. If the destination network address is not defined in the routing table, the packet will be dropped. When working with static routing or dynamic routing protocols, this is the default procedure for the IP router in most cases. Routing protocols like RIP, OSPF etc. are based on destination routing. It is also possible to use the source part of an IP data packet to make a routing decision. For this to make work, an administrator must define special route maps. Example: A route map defines to forward all data packets from the network 10.12.5.0/24 to the Ethernet interface 3, and all data packets from 10.12.6.0/24 to the next hop gateway 10.10.45.1. Rolf Augstein © 2006 All rights reserved Page 11
  • 12. The example shows no use of any destination IP addresses to make the routing decision. Routing decisions are not longer based on best paths with low metric. Routing becomes a matter of local policies. Note This is also called policy based routing. An administrative policy rules how routing decisions have to be made. When using this kind of routing, all rules for routing data traffic are defined statically in route maps. When the network becomes bigger, it can be very difficult to avoid “loosing routes somewhere in the network”. It is possible to combine destination routing with source routing within a routing node. Source routing is often used in conjunction with Quality of Service (QoS). Rolf Augstein © 2006 All rights reserved Page 12
  • 13. Interior Routing vs. Exterior Routing Interior vs. Exterior Routing Protocols IGP AS 56 EGP AS 53 IGP In larger networks it is necessary to use special routing protocols to handle the huge amount of routing information. Interior Gateway Protocols These protocols are used within an administrative area called Autonomous System (AS). Within an AS an administrator can decide with routing policy to use. Two or more Autonomous Systems can be linked together with the help of border routers. Typical routing protocols are: RIP Version 1/ 2 Routing Information Protocol OSPF Open Shortest Path First IS – IS Intermediate State – Intermediate State Cisco IGRP Interior Gateway Routing Protocol Cisco EIGRP Enhanced EIGRP Rolf Augstein © 2006 All rights reserved Page 13
  • 14. Note Autonomous Systems are identified by a 16 Bit number. This number is administrated from the Internet Assigned Numbers Authority (IANA). Two Internet RFCs discuss autonomous systems: RFC 1930 (Guidelines for creation, selection, and registration of an Autonomous System, March 1996) and RFC 0975 (Autonomous confederations, February 1986) According to RFC 1930 , "Without exception, an AS must have only one routing policy. Here routing policy refers to how the rest of the Internet makes routing decisions based on information from your AS." Exterior Gateway Protocols With an Exterior Gateway Protocol capsulated routing information within one Autonomous System is send to a second AS. The EGP connects Autonomous Systems by delivering dynamic procedures to propagate routing changes in a controlled manner. Typical routing protocols are: EGP Exterior Gateway Protocol (Old, barely used) BGP Border Gateway Protocol BGP design and configuration can be very complex. It is mostly used in some internet areas where carriers and internet providers are working together. Rolf Augstein © 2006 All rights reserved Page 14
  • 15. Routing Operation Finding the Way Routing Tables 1 Net Gateway 1a 1 Direct 2 Direct a 3 2b 4 2b 2a 2 2b Net Gateway 1 2a b 2 Direct 3 Direct 4 3c 3b 3 Net Gateway 3c 1 3b 2 3b 3 Direct c 4 Direct 4c 4 The basic idea behind routing protocols is, to send local routing information to adjacent routing nodes. All connected interfaces with a configured IP address, cause an entry in the local routing table. The routing table consists of information to reachable destination networks. Local networks are marked as “direct connected” or “local”. With routing update packets send in a given time interval, neighbor routers using the same routing protocol learn possible ways to IP networks. In the next step, all learned routes from adjacent routing nodes are sent again in the next update cycle. Rolf Augstein © 2006 All rights reserved Page 15
  • 16. If a routing node learns routes via OSPF routing, these routes are not updated by a different routing protocol like RIP. To make these protocols to interact, route redistribution is necessary. Next Hop 1 1a a I can reach network 3 and 2a 2 4 through my “Next Hop”, 2b router 2b b 3b 3 3c Next Hop: c 4c 4 Interface IP Address of the directly connected neighbor router All reachable IP destination networks are learned in the routing table. But the router has only a limited number of information sources. Example: Router “a” has only one information source, which is the adjacent router “b”. There is no information, telling router “a” that there is a third router “c”. But router “a” can reach the network “4” through router “b” as well. This information source is called the “next-hop gateway”. So after some time, a router learns all reachable networks, but is not aware of all other routers in the network. This is sometimes referred to as “routers have a flat view of the network”. Routers must have a valid route to the next-hop gateway. So always use the directly connected interface of the next-hop gateway as IP address. Rolf Augstein © 2006 All rights reserved Page 16
  • 17. Bellman Ford Algorithm The Bellman Ford algorithm is used to find the shortest way in a graph and is the basis of distance-vector routing protocols. Distance Vector Routing Protocol RT RT RT RT Interval n+2 Interval n+1 Interval n Broadcast Load Convergence Problem Metric Restrictions Distance Vector (DV) Routing The principle of DV routing is to send routing updates in a defined interval through all interfaces. These update packets use broadcast addresses, and contain information about all the reachable networks. The vector consists of the source address of the sending router. By this address receiving routers learn the address of the next-hop gateway over which the propagated networks can be reached. The distance describes the metric. In most cases this is simply the number of hops to a destination network. This number is restricted to a maximum of 15 hops. A typical DV protocol is the Routing Information Protocol (RIP). It is widely used and implemented in all UNIX and Windows Servers. Rolf Augstein © 2006 All rights reserved Page 17
  • 18. By depending on a fixed time interval to send the routing updates to all neighbors, routing information need a certain amount of time to travel through the network. This effect is called “convergence”. The use of broadcast addresses causes in the WAN part of large network some problems. The advantage of DV routing is the simple implementation and the easy way to use it in networks. Link State Algorithm Link-State algorithms are the solution for modern routing protocols. But they operate in a totally different way than the DV protocols. Link-State Routing Protocol LSA CPU Topology Database Memory RT .................. ................. ................. ................. .......... ............... SPF Algorithm Shortest Path First Tree Rolf Augstein © 2006 All rights reserved Page 18
  • 19. Link State (LS) Routing The basic concept of link-state routing is that every node receives a connectivity map of the network, in the form of a graph showing which nodes are connected to which other nodes. Each node independently calculates the best next hop from it for every possible destination in the network Each router builds a relationship with all other routers using a link-state protocol. Different roles like designated router, area router, border router etc. are assigned. Each node periodically makes up a short message, the link-state advertisement, (LSA). The LSA´s are used to identify other nodes which are directly connected and keep track of changes in routing. All information concerning other routing nodes and reachable networks are stored in the topology database. Compared to DV routing, a LS router holds more information about the entire network and does not have a flat view only. To find the best way through all the reachable destination networks, LS routing uses the algorithm SPF. Shortest Path First (SPF) A routing node uses the stored graph to calculate all paths to each other routing nodes. The paths with the best metric values are used to forward IP data packets. The result is a spanned tree with best paths to all destination networks instead of a flat view compared to DV routing. The advantage of LS routing is quick reaction to any changes in the network topology. Rolf Augstein © 2006 All rights reserved Page 19
  • 20. Process Topology Changes Link Up-Down router% Line protocol down...... or Keepalive Timer router% Line protocol up...... Entries in Routing Table .......... C 194.123.123.16 is directly connected, Ethernet0 R network 123.123.0.0 via Ethernet 0 R network 34.23.0.0 via Ethernet 0 All routes associated with interface Ethernet 0 are C 193.141.147.0 is directly connected, BRI0 not valid any longer .......... How does a routing node realize changes in the network topology? Usually, topology changes cause error states on the connected router interface. The line protocol goes down or the interface hardware fails. To control the functionality of the interfaces, the Operating System generates control packets which are sent through the interface. If the interface signals a problem, the operation state changes and all corresponding routes are effected in the routing table. Rolf Augstein © 2006 All rights reserved Page 20
  • 21. Routing Timer Update Time between Updates network unreachable .... Invalid network possibly down ..... Time after the entry is marked as “invalid” Flushing Route is erased from network 13.2.3.0 Routing table To avoid flapping interfaces and flapping routes the entire state change process uses a delay mechanism. Note The term “flapping” is often used to describe a failure condition, where i.e. an interface changes the state between up and down very often in small time intervals. This can cause a lot of problems and effect the entire network routing. An invalid timer controls when a route is marked as possibly unreachable or down. This timer is set 2 – 3 times higher than the update timer. At least two missed updates are necessary to cause a change in routing. An additionally flushing timer determines when a routing entry marked as possibly down is erased out of the routing table. Rolf Augstein © 2006 All rights reserved Page 21
  • 22. Using multiple Paths Load Balancing Route A Route B Packets are “balanced” through multiple ways More Bandwidth Advantage: Higher Availability When the routing process has two or more paths with equal metric to a destination network, it is possible to send the data packets along these routes. The data load is balanced. Some routing protocols can perform unequal cost load balancing with up to 5 different routes. Rolf Augstein © 2006 All rights reserved Page 22
  • 23. Load Balancing (cont.) Problem: Route A Route B Different Trip Times 114 ms 262 ms - Per Destination Load Balancing - Per Packet Load Balancing One of the main problems when performing load balancing exists in the different trip times of particular routes. This can cause problems for data application when the packets arrive in a different order then actually sent. Special care must be taken. Different techniques are available to solve the negative effects. Example A gateway has 2 two different paths to the headquarter network. The first session initiated is sent through the first known path, the second session is sent through the second path. The third session must use the first path and so on. This called “Per Destination Load Balancing”. When using “Per Packet Load Balancing” all packets regardless of the session ID are balanced over both paths. In this case the load on the different paths is balanced in a optimized manner. But the risk of packet delays with a higher rate of retransmissions is more likely. Rolf Augstein © 2006 All rights reserved Page 23
  • 24. Control Packet Lifetime Time-to-Live IP-Version 4 Header TTL 23 TTL 23 TTL 22 TTL 22 Decreasing Time-To-Live Counter when passing through router In the header of the IP packet, the field TTL takes care of data packets not travelling in the network for ever. Whenever a routing node forwards an IP packet, the TTL counter is decreased by one. A packet with TTL set to 0 is discarded by the router. Rolf Augstein © 2006 All rights reserved Page 24
  • 25. Routing Problems Each routing protocol has advantages and also disadvantages. There is no perfect routing protocol. An administrator must deal with the pros and the cons trying to find the best solution for his needs. Convergence Convergence Problem New Route to 194.200.1.0 194.200.1.0 194.200.1.0 194.200.1.0 194.200.1.0 194.200.1.0 300 secs 240 secs 180 secs 120 secs Next Update in 60 secs Worse case scenario A major problem with DV routing is the convergence problem. New information like changes in routing take quite some time to get to all members of the routing process. The negative effect is increasing when networks become bigger and the changes occur much more often. Rolf Augstein © 2006 All rights reserved Page 25
  • 26. Count to Infinity Count to Infinity ? Don´t worry ! I have a route to 194.200.1.0 No Route to 194.200.1.0 194.200.1.0 194.200.1.0 194.200.1.0 194.200.1.0 Worse case scenario Slow convergence causes additional problems. Routers update routing information to neighbours, even if they are the source of this information. This phenomenon is called count-to-infinity, because it leads to a ping-pong effect until the maximum value for the metric is reached. So how can one overcome this kind of effects? Rolf Augstein © 2006 All rights reserved Page 26
  • 27. Triggered Updates Solution: Triggered Updates interface down Network unreachable Metric <max. Value> Neighbor receives Update with max. Metric Overcome Convergence Problem ! Any other Changes are transmitted immediately The flow of negative information must be accelerated. Whenever a change in routing occurs, these changes are transmitted immediately to all adjacencies. If an entire network is unreachable, the update packets contain the metric value set to the maximum. Poison Reverse This technical term is used to indicate, that a packet with a higher metric or the maximum metric is set and sent along the reverse path trough the network to overcome problems like routing loops or count-to-infinity. Poison reverse is a triggered update to speed up the convergence of the routing protocol. Rolf Augstein © 2006 All rights reserved Page 27
  • 28. Interface Hold-down Update from neighbor Network 154.34 .23.0 d own Route Table er Tim Flush network 154.34.23.0 via E0 >entering hold-down for network 154.34.23.0 • Accept no further Information for Network 154.34.23.0 for a certain amount of time • Avoidance of Routing-Loops A router should not rely on information arriving on an interface that was sent out earlier over that interface. When a route is flushed out of the routing table, new update packets for a particular route from any neighbour are not accepted for some time. A router should realize which routes were propagated through the interfaces and should not accept some routes backward. Again these kinds of problems occur mainly on DV routing protocols on networks with high convergence. Rolf Augstein © 2006 All rights reserved Page 28
  • 29. Loops Routing Loops Mrs. Easy Company Intranet with 198.210.25.0 different Administrators De fau … lt R ou ute to te to … o ult R 200.200.45.0 Def a 195.22.5.0 Default Route to … Mr. Brainbox Mr. Theory Another problem coming up sometimes is a loop in the routing information table. A routing loop can be caused by a lack of communication between different routing administrators, for example. This is a very tricky problem. It looks ridicules – but it is configured very quickly. Another source for routing loops is the way DV-protocols like RIP are working as seen in previous chapter. The solution to avoid loops is the Split Horizon. Rolf Augstein © 2006 All rights reserved Page 29
  • 30. Split Horizon Split-horizon is a common solution to avoid routing loops. A cause for the route loop is that the router propagates routing information learned from a neighbour to that neighbour back. The idea of the split-horizon is not to send the routing information over the interface that has received this routing information. Split Horizon Hub and Spoke Is not propagated by RIP Propagated by RIP Dynamic Routing with RIP Can not access network 173.25.0.0 ! Network 173.25.0.0 Problem: Point-to-Multipoint Interfaces The Split Horizon problem comes up in switched wide area networks. In a switched network, one physical interface is configured with several instances of logical interfaces. The logical interfaces deal with the different IP networks. The routing process deals with the physical interface. So information learned from the way in on this physical interface is not sent out over the same physical interface. This is to avoid routing loops. So something that was designed to solve a problem now causes another problem. Administrators must be aware of the split horizon effect in point-to-multipoint interfaces to avoid routing misconfiguration. Rolf Augstein © 2006 All rights reserved Page 30
  • 31. Routing Interoperability Many administrators use more than one routing protocol in their network to manage various needs. This chapter covers how different routing protocols can configured to interact with each other. The Routing Order The Routing Preference OSPF Static Priority ? RIP OSPF 2 Static 1 RIP 3 Choice: Which routing method should be used ? Different routing protocols can be configured and activated in parallel on a router. But there is no interaction between each other. This means RIP gets all routing information for the network and a second routing protocol like OSPF calculates the best path through the some network as well. Question: So what routing paths are preferred by a data packet ? Each routing protocol including static routing methods do have an assigned priority value by default. This value is called the preference. Rolf Augstein © 2006 All rights reserved Page 31
  • 32. Note: Cisco uses the same mechanism for routing interaction. This priority value is called Administrative Distance. Working with Preference 100.0.0.0 Entries in Routing-Table Network By Metric Preference Route to 100.0.0.0 OSPF 3 5 Route to 100.0.0.0 Static 1 8 Route to 100.0.0.0 RIP 3 10 Route with best preference value If there are several routes to a destination network, the first value checked is the preference value. This means the routing procedure with the highest priority is checked first. A lower preference value means more trust for the routing source. Again, within a routing procedure like OSPF, RIP, or Static, the metric value is used to define the best path. For customization purposes the preferences can be manually configured. If an administrator wants to trust a RIP derivate route more than an OSPF route, the default preference must be changed. Different manufacturers have different specifications on the preferences/ administrative distance of the routing protocols. Rolf Augstein © 2006 All rights reserved Page 32
  • 33. The following table shows the default preferences of the routers of Quidway series produced by Huawei. In the table, a value of “0" denotes the direct route, and a value of "255" denotes any route from an untrustworthy source. Table: Default Preference Values for Quidway Series Routing Protocol Preference DIRECT 0 OSPF 10 STATIC 60 RIP 100 Internal BGP 130 OSPF AS External 150 External BGP 170 UNKNOWN 255 Except the direct route, preferences of all dynamic routing protocols can be configured manually according to the users' requirement. Rolf Augstein © 2006 All rights reserved Page 33
  • 34. Floating Static Floating Static Route 100.0.0.0 Use preference values to make static routes “interactive” ISDN Link, Serial Link, Backup 128 KB Entries in Routing-Table Network By Metric Preference Via Route to 100.0.0.0 RIP 3 10 Serial Link Route to 100.0.0.0 Static 1 20 ISDN If serial links goes down, ISDN backup is triggered by static route With the help of the preference one can make a static route more “dynamic”. By default, a static route has higher priority than all other dynamic routing procedures. One can change the behaviour, so as long as a dynamic route is present in the routing table, these routes are preferred. When for some reason the dynamic route disappears, the defined static route takes precedence. Floating static routes are often used as part of routing concepts with ISDN backup links. Rolf Augstein © 2006 All rights reserved Page 34
  • 35. Route Redistribution Route Redistribution Routing with OSPF Routing with RIP Metric 117 ? Metric 2 Metric 139 Metric 5 not compatible As mentioned earlier, each routing procedure uses proprietary metric calculations. To make them working together and exchange routing information, Route Redistribution can be used. With Route Redistribution, basically each routing procedure can be transferred in each other. There are a lot of considerations to make, when using redistribution. This entire technique is covered in detail in a later chapter. Administrators should have deeper understanding of the single routing procedures before using redistribution between them. Rolf Augstein © 2006 All rights reserved Page 35
  • 36. Redistribution Policy Define rules for redistribution Convert OSPF Routes to RIP: Starting Metric 4 Convert RIP Routes to OSPF: Starting Metric 9 The basic principle with route redistribution consists in the choice for special routing nodes in the network, where redistribution should be established. Example: A set of definitions rule the way, a RIP route is converted and transferred in an OSPF route and vice versa. OSPF Metric 230 is converted to RIP Metric 4 RIP Metric 3 is converted to OSPF Metric 9 All metric conversions must be set with care, so the entire routing information context makes sense. Also, the choice of the position of the router in the network redistributing routes is relevant. Rolf Augstein © 2006 All rights reserved Page 36
  • 37. Routing Design A structured network design is the fundament for implementing a useful routing strategy. Without a proper IP addressing, there is no way for scalable and stable networks. Routing Hierarchy Building Internet Areas, Domains, AS Core Edge/ Convergence Access In larger networks with structured network design, some routers will take special control and handling of routing updates. To take advantage of different routing mechanisms it is very important to have a well administrated IP address scheme. Smart Router: From the routing perspective, some routers are smarter than others. Because of the routing information they hold in their routing table, some routers may have a more detailed knowledge about the network. This is common technique to control the amount of routing information. Small routers in the access zone do not need all information about the entire network. Rolf Augstein © 2006 All rights reserved Page 37
  • 38. A good IP address plan implemented in a well-designed network has the following characteristics: • Scalability Allows for large increases in the number of supported sites • Predictability Exhibits predictable behavior and performance • Flexibility Minimizes the impact of routers, additions, changes, or removals Rolf Augstein © 2006 All rights reserved Page 38
  • 39. The Prefix Prefix Prefix Host “Classfull” Routes Class A 10.0.0.0/8 10.0.0.0 255.0.0.0 Class B 129.12.0.0/16 129.12.0.0 255.255.0.0 Class C 201.12.23.0/24 201.12.23.0 255.255.255.0 “Classless” Routes Class C 201.12.112.0/21 201.12.112.0 255.255.248.0 For routing purposes, an IP address without a given subnet mask is “worthless”. To make routing decisions the subnet mask must always be considered. Instead of using the subnet mask in the dotted decimal format, a more convenient format is used. The Prefix points out, how many bits within the 32 bits of the IP address are used as the network part. So a prefix of 20 bits for an IP address like 144.37.99.34 means, you deal with a class B network 144.37.0.0 performing 4 bit subnetting. Rolf Augstein © 2006 All rights reserved Page 39
  • 40. Summarize Routes Route Summarization Prefix Host Subnetting - Gain more routable networks - Search common network bits for summarization Prefix Host Summarization The process of divide a network in smaller sub-networks is done by shifting the network bits to the right. (see TCP/IP fundamentals). When dealing with large networks, it is important to minimize the amount of routing information. Less routing information means less routing update traffic and less RAM (memory) needed in the router. So the process of summarize many sub-networks to one network is called Route Summarization. This is done by shifting the network bits to the left. Rolf Augstein © 2006 All rights reserved Page 40
  • 41. IP Address Management Route Summarization Route Aggregation 132.17.25.0 132.1 7.0.0 /16 132.17.26.0 Only 1 update necessary 132.17.27.0 132.17.28.0 132.17.29.0 IP subnetworks are auto-summarized based on Class A, B, C addresses By default, most routers perform auto summarization for class A, B, or C networks. Instead of propagating up to 254 subnets of the network 132.17.0.0 (132.17.1.0 to 132.17.254.0) the summarized route 132.17.0.0/16 is used. This means an enormous improvement for the amount of routing traffic sent to the neighbour router. Note: Sometimes the term Route Aggregation is used. An aggregate route includes different sub-networks by using appropriate subnet masks. Rolf Augstein © 2006 All rights reserved Page 41
  • 42. Relevant Bits IP Address: 10.20.35.5/ 16 10.20.35.5 00001001 00010100 00100011 00000101 16 Bits Prefix: Marks the relevant bits for all routing decisions 10.20.35.5 00001001 00010100 00100011 00000101 Logical “AND” 255.255.0.0 11111111 11111111 00000000 00000000 What the Router does ! Network: 00001001 00010100 00000000 00000000 Bits to care Don´t care Routing nodes need the IP address and the corresponding subnet mask to make routing decisions. Each interface needs this information as part of the configuration. With the help of the subnet mask and the logical “AND” operation, it is a simple process to read out the network part of the IP address. The prefix bits define the network relevant bits within an IP address. This is the reason for sending the prefix in each routing update when using routing protocols like OSPF or RIP version 2, so different subnet masks can be used within a single IP network. The older RIP version 1 is not capable of using different subnet masks in one class A, B, or C network. RIP updates are not aware of network prefix. Rolf Augstein © 2006 All rights reserved Page 42
  • 43. Sub-Subnet Host A IP Address: 10.20.35.5/ 16 10.20.35.5/16 00001001 00010100 00100011 00000101 Subnet 10.20 10.20.35.5/19 00001001 00010100 0010 0011 00000101 Subnet 10.20.32 Host B IP Address: 10.20.35.5/ 19 Without the subnet mask information, it is not possible to determine the location of a given host in the network. VLSM is like using an additional subnet for a “main subnet”. A “sub subnet” describes how many subnets are used within a defined subnet. Working with VLSM is simple math, but can be complex in real live. Interesting: Only routing nodes with appropriate routing procedures must be “aware” of VLSM. So, not all routing protocols can be used. End systems like hosts or servers do not have to deal with VLSM. They just have to be configured with a proper IP address and mask. Rolf Augstein © 2006 All rights reserved Page 43
  • 44. VLSM Routing VLSM 172.16.56.0/24 172.16.13.4/30 172.16.11.0/24 172.16.13.8/30 172.16.0.0/16 Optimization with use of various prefix subnets Variable Length Subnet Mask is often used, to optimize the address space for a given class A, B, or C network. There are lots of small networks with few hosts. Using large subnets like 8 bits prefix actually wastes a lot of address space. Worse case is a PPP link with the need of two valid IP addresses only. With a 8 bits subnet, there are 252 wasted IP addresses ! Rolf Augstein © 2006 All rights reserved Page 44
  • 45. Variable-Length Subnet Mask 172.16.0.0/24 172.16.1.0 Aggregate Route 172.16.2.0 254 . . . . 172.16.14.0/30 Subnets 172.16.14.0 . . . . 172.16.254.0 172.16.14.4 172.16.14.8 62 . . . . Subnets 172.16.14.252 Use one subnet to split into smaller VLSM subnets A proven way of using VLSM is, to take a certain subnet out of the group of available subnets. Apply the new subnet mask i.e. 30 bits, so 62 new subnets are addressable. Each new subnet can address two hosts. The benefit is the gain of new smaller routable networks, which can be used to address PPP links. Note: VLSM does not mean an increasing of IP addresses at all. As a matter of fact, lots of addresses are lost because of broadcasts and network addresses. Rolf Augstein © 2006 All rights reserved Page 45
  • 46. Table: Prefix Calculation CIDR Netmask Hosts / Class Typical usage subnet /8 255.0.0.0 16777216 A Largest block allocation made by IANA /9 255.128.0.0 8388608 /10 255.192.0.0 4194304 /11 255.224.0.0 2097152 /12 255.240.0.0 1048576 /13 255.248.0.0 524288 /14 255.252.0.0 262144 /15 255.254.0.0 131072 /16 255.255.0.0 65536 B /17 255.255.128.0 32768 ISP / large business /18 255.255.192.0 16384 ISP / large business /19 255.255.224.0 8192 ISP / large business /20 255.255.240.0 4096 Small ISP / large business /21 255.255.248.0 2048 Small ISP / large business /22 255.255.252.0 1024 /23 255.255.254.0 512 /24 255.255.255.0 256 C Large LAN /25 255.255.255.128 128 Large LAN /26 255.255.255.192 64 Small LAN /27 255.255.255.224 32 Small LAN /28 255.255.255.240 16 Small LAN /29 255.255.255.248 8 /30 255.255.255.252 4 "Glue network" (point to point links) /31 255.255.255.254 2 "Useless Network", proposed for point to point links (RFC 3021) /32 255.255.255.255 1 Host route Rolf Augstein © 2006 All rights reserved Page 46
  • 47. Classless Routing Contains block of: Defined Summary Route: 200.16.168.0 200.16.168.0/21 200.16.169.0 200.16.170.0 200.16.171.0 200.16.172.0 200.16.173.0 200.16.174.0 200.16.175.0 CIDR: Classless Inter-Domain Routing The IP address space was divided into three main network classes, where each class had a fixed network size. The class, the length of the subnet mask and the number of hosts on the network, could always be determined from the most significant bits of the IP address. Without any other way of specifying the length of a subnet mask, routing protocols necessarily used the class of the IP address specified in route advertisements to determine the size of the routing prefixes to be set up in the routing tables. CIDR uses VLSM to allocate IP addresses to subnets according to individual needs. Thus the network/host division can occur at any bit boundary in the address. The process can be recursive, with a portion of the address space being further divided into even smaller portions, through the use of masks which cover more bits. Because the normal class distinctions are ignored, the new system is called classless routing. Rolf Augstein © 2006 All rights reserved Page 47
  • 48. Prefix aggregation Another benefit of CIDR is the possibility of routing prefix aggregation. For example, sixteen contiguous /24 networks could now be aggregated together, and advertised to the outside world as a single /20 route (if the first 20 bits of their network addresses match). Two contiguous /20s could then be aggregated to a /19, and so forth. This allows a significant reduction in the number of routes that had to be advertised over the Internet, preventing 'routing table explosion' from overwhelming routers, and stopping the Internet from expanding further. When dealing with aggregate routes within the internet the term “Supernet” is used sometimes. These kinds of routing mechanisms are part of BGP routing. The Border Gateway Protocol is discussed more detailed in a later module. CIDR is described in: RFC 1519 (http://www.ietf.org/rfc/rfc1519.txt) Classless Inter-Domain Routing (CIDR): an Address Assignment and Aggregation Strategy. RFC 1518 (http://www.ietf.org/rfc/rfc1518.txt) Architecture for IP Address Allocation with CIDR Rolf Augstein © 2006 All rights reserved Page 48
  • 49. Discontinuous Use of Subnets Routing with RIP, Auto-summarization 155.10.34.0/24 198.23.24.0/24 155.10.35.0/24 ? Oh fine: I have two routes by RIP to 155.10.0.0 Another interesting effect comes up with the discontinuous use of a class A, B, or C network, which is important to understand for routing administrators. Because routers perform auto-summarization on IP network address borders, the above situation arises for a router between two networks using discontinuous IP address spaces. From the routing perspective, there are two paths to the network 155.10.0.0 – with fatal consequences ! It is not recommended to split IP networks and use them on different discontinuous locations. To solve the above problem, auto-summary must be disabled on both routers. But then, too many routes are propagated through the network cloud, which could lead to additional problems. Rolf Augstein © 2006 All rights reserved Page 49
  • 50. Prefix Matching Priority 192.16.3.33 / 32 Host 192.16.3.32 / 27 Subnet 192.16.3.0 / 24 Net 192.16.0.0 / 16 Block Network 0.0.0.0 / 0 Default Rule: “best prefix matches” When using subnetting and VLSM in a network, the routing table has various entries for a network with different prefix lengths. Longest prefix match or best prefix match refers to an algorithm used to decide for the best routing entry. Because each entry of a routing table may specify a range of addresses, one destination address may match more than another routing table entry. The most specific table entry, this means the one with smallest host address range, is called the longest prefix match. Rolf Augstein © 2006 All rights reserved Page 50
  • 51. Module Review Summary Static routing is still as important as adaptive routing protocols. Adaptive routing protocols are divided in Distance-Vector and Link-State protocols. Routing decisions are based on preferences and metric calculations. Network administrators must be aware of different routing problems like Split Horizon, Convergence, Loops, or other effects depending on the routing protocol. Different routing protocols can interact with the help of routing redistribution. Network design and appropriate IP addressing schemes are important for fast and stable routing. The ability for route summarization and aggregation is the key for adaptive routing in larger networks. Rolf Augstein © 2006 All rights reserved Page 51
  • 52. Review Question 1. Outline the difference between metric and preference? 2. What are common problems of D-V routing protocols? 3. Build a small table and outline the advantages and disadvantages of L-S routing protocols. 4. What is the meaning of an adjacent router? Rolf Augstein © 2006 All rights reserved Page 52
  • 53. 5. What is meant by a “Floating Static Route” ? 6. Describe the rule “best prefix matches” and the relevancy to routing protocols. 7. What kind of topology is Hub-and-Spoke? 8. Describe the problems arising with a slow convergence. 9. What is the preference for a direct connected network ? Why ? 10. What is the meaning of asymmetric routing ? Rolf Augstein © 2006 All rights reserved Page 53