1. Saeed Darvish Pazoki – MCSE, CCNA Abstracted From: Cisco Press – ICND 2 – 8 Routing Protocol Theory 1

2. This chapter covers the following subjects: Dynamic Routing Protocol Overview: This section introduces the core concepts behind how routing protocols work and many terms related to routing protocols. Distance Vector Routing Protocol Features: This section explains how distance vector routing protocols work, focusing on the loop-avoidance features. Link-State Routing Protocol Features: This section explains how link-state routing protocols work, using OSPF as a specific example. 2

3. Routing Protocol Theory Routers add IP routes to their routing tables using three methods: Connected routes Static routes Routes learned through dynamic routing protocols 3

4. Routing Protocol Theory Routing and Routed protocols Routing protocol: A set of messages, rules, and algorithms used by routers for the overall purpose of learning routes. This process includes the exchange and analysis of routing information. Each router chooses the best route to each subnet (path selection) and finally places those best routes in its IP routing table. Examples include RIP, EIGRP, OSPF, and BGP. Routed protocol and routable protocol: Both terms refer to a protocol that defines a packet structure and logical addressing, allowing routers to forward or route the packets. Routers forward, or route, packets defined by routed and routable protocols. Examples include IP and IPX (a part of the Novell NetWare protocol model). 4

5. Routing Protocol Theory Routing Protocol Functions 1. Learn routing information about IP subnets from other neighboring routers. 2. Advertise routing information about IP subnets to other neighboring routers. 3. If more than one possible route exists to reach one subnet, pick the best route based on a metric. 4. If the network topology changes—for example, a link fails— react by advertising that some routes have failed, and pick a new currently best route. (This process is called convergence.) 5

6. Routing Protocol Theory 6

7. Interior and Exterior Routing Protocols IP routing protocols fall into one of two major categories: 1. Interior Gateway Protocols (IGP) 2. Exterior Gateway Protocols (EGP). The definitions of each are as follows: IGP: A routing protocol that was designed and intended for use inside a single autonomous system (AS) EGP: A routing protocol that was designed and intended for use between different autonomous systems 7

8. Routing Protocol Theory Autonomous System (AS) An AS is an internetwork under the administrative control of a single organization. Some routing protocols work best inside a single AS by design, so these routing protocols are called IGPs. Conversely, routing protocols designed to exchange routes between routers in different autonomous systems are called EGPs. 8

9. Routing Protocol Theory Autonomous System (AS) Each AS can be assigned a number called (unsurprisingly) an AS number (ASN). Like public IP addresses, the Internet Corporation for Assigned Network Numbers (ICANN, http://www.icann.org) controls the worldwide rights to assigning ASNs. It delegates that authority to other organizations around the world, typically to the same organizations that assign public IP addresses. For example, in North America, the American Registry for Internet Numbers (ARIN, http://www.arin.net/) assigns public IP address ranges and ASNs. 9

10. Routing Protocol Theory Autonomous System (AS) 10

11. Routing Protocol Theory Comparing IGPs Today, there is no real choice of what EGP to use: you simply use BGP. However, when choosing an IGP to use inside a single organization, you have several choices. The most reasonable choices today are RIP-2, EIGRP, and OSPF. 11

12. Routing Protocol Theory IGP Routing Protocol Algorithms A routing protocol’s underlying algorithm determines how the routing protocol does its job. Three main branches of routing protocol algorithms exist for IGP routing protocols: Distance vector Link-state Balanced hybrid (sometimes called enhanced distance vector) Distance vector protocols suffer from slow convergence and they are potential for routing loops! 12

13. Routing Protocol Theory IGP Routing Protocol Algorithms Metrics Routing protocols choose the best route to reach a subnet by choosing the route with the lowest metric. 13

14. Routing Protocol Theory IGP Routing Protocol Algorithms 14

15. Routing Protocol Theory IGP Routing Protocol Algorithms 15

16. Routing Protocol Theory IGP Routing Protocol Algorithms 16

17. Routing Protocol Theory Administrative Distance Many companies and organizations use a single routing protocol. However, in some cases, a company needs to use multiple routing protocols. When a single routing protocol learns multiple routes to the same subnet, the metric tells it which route is best. However, when two different routing protocols learn routes to the same subnet, because each routing protocol’s metric is based on different information, IOS cannot compare the metrics. When IOS must choose between routes learned using different routing protocols, IOS uses a concept called administrative distance. 17

18. Routing Protocol Theory Administrative Distance 18

19. Routing Protocol Theory The Concept of a Distance and a Vector The term distance vector describes what a router knows about each route. At the end of the process, when a router learns about a route to a subnet, all the router knows is some measurement of distance (the metric) and the next-hop router and outgoing interface to use for that route (a vector, or direction). 19

20. Routing Protocol Theory The Concept of a Distance and a Vector 20

21. Routing Protocol Theory The Concept of a Distance and a Vector 21

22. Routing Protocol Theory Distance Vector Operation in a Stable Network Distance vector routing protocols send periodic full routing updates. 22

23. Routing Protocol Theory Distance Vector Loop Prevention Route Poisoning When a route fails, distance vector routing protocols risk causing routing loops until every router in the internetwork believes and knows that the original route has failed. Distance vector protocols spread the bad news about a route failure by poisoning the route. Route poisoning refers to the practice of advertising a route, but with a special metric value called infinity. Simply put, routers consider routes advertised with an infinite metric to have failed. RIP defines infinity as 16. 23

24. Routing Protocol Theory Distance Vector Loop Prevention Route Poisoning 24

25. Routing Protocol Theory Distance Vector Loop Prevention Problem: Counting to Infinity over a Single Link Distance vector routing protocols risk causing routing loops during the time between when the first router realizes a route has failed until all the routers know that the route has failed. Counting to infinity causes two related problems. Several of the distance vector loop prevention features focus on preventing these problems: Packets may loop around the internetwork while the routers count to infinity, with the bandwidth consumed by the looping packets crippling an internetwork. The counting-to-infinity process may take several minutes, meaning that the looping could cause users to believe that the network has failed. 25

26. Routing Protocol Theory Distance Vector Loop Prevention Problem: Counting to Infinity over a Single Link 26

27. Routing Protocol Theory Distance Vector Loop Prevention Split Horizon Split horizon is defined as follows: In routing updates sent out interface X, do not include routing information about routes that refer to interface X as the outgoing interface. 27

28. Routing Protocol Theory Distance Vector Loop Prevention Split Horizon 28

29. Routing Protocol Theory Distance Vector Loop Prevention Poison Reverse and Triggered Updates Distance vector protocols can attack the counting-to-infinity problem by ensuring that every router learns that the route has failed, through every means possible, as quickly as possible. The next two loop-prevention features do just that and are defined as follows: Triggered update: When a route fails, do not wait for the next periodic update. Instead, send an immediate triggered update listing the poisoned route. Poison reverse: When learning of a failed route, suspend split- horizon rules for that route, and advertise a poisoned route. 29

30. Routing Protocol Theory Distance Vector Loop Prevention Poison Reverse and Triggered Updates 30

31. Routing Protocol Theory Distance Vector Loop Prevention Problem: Counting to Infinity in a Redundant Network Split horizon prevents the counting-to-infinity problem from occurring between two routers. However, with redundant paths in an internetwork, which is true of most internetworks today, split horizon alone does not prevent the counting-to-infinity problem. 31

32. Routing Protocol Theory Distance Vector Loop Prevention Problem: Counting to Infinity in a Redundant Network 32

33. Routing Protocol Theory Distance Vector Loop Prevention The Holddown Process and Holddown Timer Holddown, prevents the looping and counting-to-infinity problem. Distance vector protocols use holddown to specifically prevent the loops created by the counting-to-infinity problems that occur in redundant internetworks. The holddown process can be summarized as follows: After hearing a poisoned route, start a holddown timer for that one route. Until the timer expires, do not believe any other routing information about the failed route, because believing that information may cause a routing loop. However, information learned from the neighbor that originally advertised the working route can be believed before the holddown timer expires. 33

34. Routing Protocol Theory Distance Vector Loop Prevention The Holddown Process and Holddown Timer 34

35. Routing Protocol Theory Distance Vector Summary During periods of stability, routers send periodic full routing updates based on a short update timer (the RIP default is 30 seconds). The updates list all known routes except the routes omitted because of split-horizon rules. When changes occur that cause a route to fail, the router that notices the failure reacts by immediately sending triggered partial updates, listing only the newly poisoned (failed) routes, with an infinite metric. Other routers that hear the poisoned route also send triggered partial updates, poisoning the failed route. Routers suspend split-horizon rules for the failed route by sending a poison reverse route back toward the router from which the poisoned route was learned. All routers place the route in holddown state and start a holddown timer for that route after learning that the route has failed. Each router ignores all new information about this route until the holddown timer expires, unless that information comes from the same router that originally advertised the good route to that subnet. 35

36. Routing Protocol Theory Link-State Routing Protocol Features Building the Same LSDB on Every Router Routers using link-state routing protocols need to collectively advertise practically every detail about the internetwork to all the other routers. At the end of the process, called flooding, every router in the internetwork has the exact same information about the internetwork. Routers use this information, stored in RAM inside a data structure called the link-state database (LSDB), to perform the other major link-state process to calculate the currently best routes to each subnet. 36

37. Routing Protocol Theory Link-State Routing Protocol Features Open Shortest Path First (OSPF), the most popular link-state IP routing protocol, advertises information in routing update messages of various types, with the updates containing information called link-state advertisements (LSA). LSAs come in many forms, including the following two main types: Router LSA: Includes a number to identify the router (router ID), the router’s interface IP addresses and masks, the state (up or down) of each interface, and the cost (metric) associated with the interface. Link LSA: Identifies each link (subnet) and the routers that are attached to that link. It also identifies the link’s state (up or down). 37

38. Routing Protocol Theory Link-State Routing Protocol Features 38

39. Routing Protocol Theory Link-State Routing Protocol Features Applying Dijkstra SPF Math to Find the Best Routes The link-state flooding process results in every router having an identical copy of the LSDB in memory, but the flooding process alone does not cause a router to learn what routes to add to the IP routing table. Link-state protocols must use another major part of the link-state algorithm to find and add routes to the IP routing table—routes that list a subnet number and mask, an outgoing interface, and a next-hop router IP address. This process uses something called the Dijkstra Shortest Path First (SPF) algorithm 39

40. Routing Protocol Theory Link-State Routing Protocol Features Applying Dijkstra SPF Math to Find the Best Routes The SPF algorithm can be compared to how humans think when taking a trip using a road map. If several routes look similar in length, you may decide to take a longer route if the roads are highways rather than country roads. Someone else may own the same map, but they might be starting from a different location, and going to a different location, so they may choose a totally different route. In the analogy, the LSDB works like the map, and the SPF algorithm works like the person reading the map. 40

41. Routing Protocol Theory Link-State Routing Protocol Features 41

42. Routing Protocol Theory Summary and Comparisons to Distance Vector Protocols Link-state routing protocols provide fast convergence, which is probably the most important feature of a routing protocol, with built-in loop avoidance. Link-state routing protocols do not need to use the large variety of loop-avoidance features used by distance vector protocols, which in itself greatly reduces the convergence time. 42

43. Routing Protocol Theory Summary and Comparisons to Distance Vector Protocols The main features of a link-state routing protocol are as follows: All routers learn the same detailed information about all routers and subnets in the internetwork. The individual pieces of topology information are called LSAs. All LSAs are stored in RAM in a data structure called the link-state database (LSDB). Routers flood LSAs when 1) they are created 2) on a regular (but long) time interval if the LSAs do not change over time 3) immediately when an LSA changes. The LSDB does not contain routes, but it does contain information that can be processed by the Dijkstra SPF algorithm to find a router’s best route to reach each subnet. Each router runs the SPF algorithm, with the LSDB as input, resulting in the best (lowest-cost/lowest-metric) routes being added to the IP routing table. Link-state protocols converge quickly by immediately re-flooding changed LSAs and rerunning the SPF algorithm on each router. 43

44. Routing Protocol Theory The following list summarizes some of the key comparison points for different routing protocols, comparing the strengths of the underlying algorithms: Convergence: Link-state protocols converge much more quickly. CPU and RAM: Link-state protocols consume much more CPU and memory than distance vector protocols, although with proper design, this disadvantage can be reduced. Avoiding routing loops: Link-state protocols inherently avoid loops, whereas distance vector protocols require many additional features (for example, split horizon). Design effort: Distance vector protocols do not require much planning, whereas link-state protocols require much more planning and design effort, particularly in larger networks. Configuration: Distance vector protocols typically require less configuration, particularly when the link-state protocol requires the use of more-advanced features. 44