Chapter 17 Implementing Redistribution and Controlling Routing Updates
Chapter 18 Controlling Network Traffic with Route Maps and Policy-Based Routing
Part VI covers the following Cisco BSCI exam topics:
■ Identify the steps to select and configure the different ways to control routing update traffic
■ Identify the steps to configure policy-based routing using route maps
■ Identify the steps to configure router redistribution in a network
■ Explain the use of redistribution between BGP and Interior Gateway Protocols (IGPs)
■ Identify the steps to verify route redistribution
■ Interpret the output of various show and debug commands to determine the cause of route selection errors and configuration problems
This chapter covers the following topics, which you need to understand to
pass the CCNP/CCDP/CCIP
■ Understanding the fundamentals of redistribution
■ Identifying the issues with redistribution
■ Understanding the routing decisions that affect redistribution
■ Controlling routing updates during redistribution
■ Conﬁguring redistribution
■ Conﬁguration commands to control routing updates in redistribution
■ Controlling routing updates with ﬁltering
■ Verifying, maintaining, and troubleshooting the implementation of redistribution and ﬁltering
Implementing Redistribution and Controlling Routing Updates
The topics in this chapter deal with the trafﬁc generated by routing updates in terms of both the network resources that they use and the information contained within them. This covers two different but related areas, redistribution and ﬁltering. The network overhead involved in routing updates has already been dealt with in other chapters. It keeps recurring as a theme because all network trafﬁc directly inﬂuences the network’s capability to scale or to grow.
The information propagated through the network is complex when dealing with one routing protocol. When multiple protocols have to share information (through redistribution) so that the larger network can see every route available within the autonomous system, the information ﬂow must be controlled and managed very closely with ﬁltering and other solutions.
This chapter deals with the need for redistribution, which increases the network overhead, and ﬁltering, which is used to reduce overhead. The chapter explains the design issues that might affect the conﬁguration, followed by conﬁguration examples of route redistribution and ﬁltering.
“Do I Know This Already?” Quiz
The purpose of the “Do I Know This Already?” quiz is to help you decide what parts of this chapter to use. If you already intend to read the entire chapter, you do not necessarily need to answer these questions now.
The 14-question quiz, derived from the major sections in the “Foundation Topics” portion of the chapter, helps you to determine how to spend your limited study time.
Table 17-1 outlines the major topics discussed in this chapter and the “Do I Know This Already?” quiz questions that correspond to those topics.
Table 17-1 “Do I Know This Already?” Foundation Topics Section-to-Question Mapping
CAUTION The goal of self-assessment is to gauge your mastery of the topics in this chapter. If you do not know the answer to a question or are only partially sure of the answer, you should mark this question wrong for purposes of the self-assessment. Giving yourself credit for an answer you correctly guess skews your self-assessment results and might provide you with a false sense of security.
1. Which of the following are reasons to run multiple routing protocols?
a. Applications requiring UNIX
b. Smaller routing domains speed up convergence
c. Political enclaves
d. Smaller routing domains ensure a more stable network
2. EIGRP automatically redistributes into which routing protocols?
a. IGRP running the same AS number
b. IGRP with any process ID
c. EIGRP running the same AS number
d. EIGRP with any process ID
3. What do the letters SIN represent?
a. Simple Internetwork
b. Ships in the night
c. Structured IP Network
d. Single IP Network
4. How many IP routing tables can be held on a router within a typical organization?
b. One per routing protocol
5. The problems experienced as a result of multiple routing processes and their redistribution include which of the following?
a. Suboptimal path
b. Loss of Hello packets
c. Routing loops
d. Continuous LSA propagation
6. What action is taken if no seed or default metric is conﬁgured for OSPF when redistributing EIGRP?
a. The route is not entered into the routing table.
b. The route is entered with a cost of 0.
c. The route is read into OSPF with a cost of 20 (type E2).
d. The route is entered with a cost of 20 (type 1).
7. What techniques can be employed to avoid redistribution problems?
a. Distribute lists
b. Change administrative distance
c. Ensuring the default metric is set to 0
d. Redistributing on all border routers in both directions
8. What is the purpose of distribute lists?
a. Determine the administrative distance of a distributed routing protocol
b. Identify which interfaces will send updates
c. Determine which networks are sent in updates
d. Determine which networks are accepted into the routing table
9. Where are distribute lists deﬁned?
a. At the interface
b. Under the routing process
c. At the router level
d. At the executive prompt
10. Which command is used to establish the default or seed metric for EIGRP?
a. default-metric 5
b. metric bandwidth delay reliability loading mtu
c. default-metric bandwidth delay reliability loading mtu eigrp
d. default-metric bandwidth delay reliability loading mtu
11. Which command is used to conﬁgure the administrative distance?
a. administrative distance
c. ip default-distance
d. ip administrative distance
12. Why might it be necessary to control the routing updates?
b. Prevention of routing loops
c. Scaling the network
d. Preserving the metric
13. All of the following statements are true; however, which of the following actions is taken ﬁrst?
a. Do not advertise the route if it is matched by a deny statement.
b. If no match is found in the distribute list, the implicit deny any at the end of the access list will cause the update to be dropped.
c. If a ﬁlter is present, the router examines the access list to see if there is a match on any of the networks in the routing update.
d. Advertise the route if matched by a permit statement.
14. Which commands could be used to verify and troubleshoot a network that is redistributed?
a. show ip protocol
b. show ip route
c. show ip route routing-protocol
d. show redistributed
The answers to this quiz are found in Appendix A, “Answers to Chapter ‘Do I Know This Already?’ Quizzes and Q&A Sections.” The suggested choices for your next step are as follows:
■ 7 or less overall score —Read the entire chapter. This includes the “Foundation Topics” and “Foundation Summary” sections, the “Q&A” section, and the “Scenarios” at the end of the chapter.
■ 8–11 overall score —Begin with the “Foundation Summary” section, and then go to the “Q&A” section and the “Scenarios” at the end of the chapter. If you have trouble with these exercises, read the appropriate sections in “Foundation Topics.”
■ 12 or more overall score —If you want more review on these topics, skip to the “Foundation Summary” section, and then go to the “Q&A” section and the “Scenarios” at the end of the chapter. Otherwise, move to the next chapter.
Understanding the Fundamentals of Redistribution
It is rare to ﬁnd just one routing protocol running within an organization. If the organization is running multiple routing protocols, you need to ﬁnd some way of passing the networks learned by one routing protocol into another so that every workstation can reach every other workstation. This process is called redistribution.
Redistribution is used when a router is receiving information about remote networks from various sources. Although all the networks are entered into the routing table and routing decisions are made on all the networks present in the table, a routing protocol propagates only those networks that it has learned through its own process. When there is no sharing of network information between the routing processes, it is referred to as ships in the night (SIN) routing.
Redistribution is often necessary within a network, if only as a transitional implementation. Nonetheless, it should not be thought of as a quick and easy solution. Although route redistribution is often a lifesaver for your network, it is fraught with complexity. Understanding the operation of the processes that you have implemented and how this inﬂuences your network is crucial. This chapter focuses on the main topics dealing with the implementation of redistribution.
Although an organization might have many routing protocols running within its autonomous system, each interior routing protocol sees itself as the only interior routing protocol within the autonomous system. When an interior routing protocol such as EIGRP has routes redistributed into its routing process, it assumes that these routes are from another autonomous system and are therefore external routes. This affects the route selection made by the routing process, and EIGRP prefers the interior routes.
The exterior routing protocols see the organization as the autonomous system that connects to the Internet or a service provider.
In Figure 17-1, the routing table for Router B has entries from RIP and OSPF. There are no entries for EIGRP because this is a single network directly connected to the router. You can see that the RIP updates sent out the interfaces do not include networks from OSPF. There are no entries for EIGRP.
Furthermore, Router C has only connected routes in the routing table. This is because, although EIGRP has been conﬁgured, Router C is a stub router. When the other interfaces are conﬁgured with addresses and the rest of the EIGRP network is connected to Router C, the network will be populated
with EIGRP routes, which it will propagate to Router B. If redistribution is then implemented, the entire network will be available to everyone.
Figure 17-1 Routing Updates Without Using Redistribution
Redistribution can occur only between processes routing the same Layer 3 protocol. So, for example, OSPF, RIP, IGRP, and EIGRP can redistribute routing updates among themselves because they all support the same TCP/IP stack and share the same routing table. However, there can be no network redistribution between AppleTalk and IPX.
Some routing protocols automatically exchange networks, although others require some level of conﬁguration. Table 17-2 shows the subtleties of automatic redistribution.
Table 17-2 Automatic Redistribution Between Routing Protocols
Figure 17-2 illustrates redistribution within an organization.
The main reasons for multiple protocols existing within an organization are as follows:
■ The organization is transitioning from one routing protocol to another because there is a need for a more sophisticated protocol.
■ Historically, the organization was a series of small network domains. The company has plans to transition to a single routing protocol in the future.
■ Some departments might have host-based solutions that require different protocols. For example, some UNIX hosts use RIP to discover gateways.
■ Often after a merger or a takeover, it takes planning, strategy, and careful analysis to determine the best overall design for the network.
Figure 17-2 Autonomous Systems Within an Organization
■ Politically, there are ideological differences among the different network administrators, which until now have not been resolved.
■ In a very large environment, the various domains might have different requirements, making a single solution inefﬁcient. A clear example is in the case of a large multinational corporation, where EIGRP is the protocol used at the access and distribution layers, but BGP is the protocol connecting the core.
Understanding the Routing Decisions That Affect Redistribution
When embarking on running multiple routing protocols within your network and making one cohesive whole network, redistribution is the answer, but only after careful consideration has been given to the problems that might arise. In order to do this, you need to consider brieﬂy the routing protocol operation, in particular how a path is selected to go into the routing table. For a detailed discussion on routing tables, refer to Chapter 1, “IP Routing Principles.” Path selection is dealt with in depth in Chapter 4, “IP Distance Vector Routing Principles.”
Routing Metrics and Redistribution
There are many routing protocols for IP, and each routing protocol uses a different metric. If the different protocols want to share information through redistribution, the conﬁguration must translate the metrics. The conﬁguration commands are dealt with in the section “Conﬁguring Redistribution,” later in this chapter.
Problems arise when the metrics are redistributed without additional conﬁguration. The metric has no point of reference in the new routing protocol; for example, RIP would be bafﬂed by the metric presented as 786, when expecting a hop count between 0–15. In accepting the new networks, the receiving process must have a starting point, or seed metric, in order to calculate the metric for the routing protocol.
The seed metric is assigned to all the routes received into a process through redistribution. The metric is incremented from that point on as the networks propagate throughout the new routing domain.
There are defaults for the seed metrics, but depending on the routing protocol, the default might prevent the route from entering the routing table. The seed metrics are as deﬁned in Table 17-3.
Table 17-3 Default Seed Metrics
Remember, the metric is the main method of route selection within a routing protocol. Therefore, it is necessary to deﬁne a default seed metric for the networks accepted from the other routing protocol.
Path Selection Between Routing Protocols
Now that route selection within a routing protocol has been explained, this section discusses path selection between routing protocols when more than one routing protocol is running on your network. If the protocols have paths to the same remote destination network, the routing process must decide which path to enter into the routing table. Because the metrics differ between the protocols, selection based on the metric is ruled out as a solution. Instead, another method was devised to solve the problem, the administrative distance, as discussed in Chapter 4.
The distinction between the two selection processes is simple: Administrative distance determines between IP routing protocols, and the metric chooses between paths from one routing protocol.
Administrative distance and metrics appear to solve all your problems, until you start to redistribute the information between routing protocols, and the routing process becomes confused as to from where the information came. When the carefully determined rules of selection become tangled, suboptimal routing decisions and routing loops result.
It is therefore important to consider the following rules when redistributing between IP routing protocols:
■ If more than one routing protocol is running on a router, the routing table will place the route with the best administrative distance into the routing table.
■ In order to be redistributed, the route must exist in the routing table under the ownership of the routing protocol that is being redistributed. Thus, if RIP is being redistributed into EIGRP, the routing table must have an entry for the RIP network.
■ When a route is redistributed, it inherits the default administrative distance of the new routing protocol.
■ When a route is redistributed, it is considered as an external route to the new routing protocol. For EIGRP and BGP, this means it will inherit the administrative distance of an external route to the new routing protocol. OSPF tracks the route as external and chooses internal routes ﬁrst.
It is clear that redistribution is not the optimum network design. The simpler and more straightforward the design, the better managed and more stable the network, with fewer errors and faster convergence. Therefore, a hierarchical IP addressing scheme designed to allow continued network growth, combined with a single IP routing protocol that has the scope to support growth, results in a strong, reliable, and fast network. However, it is rare to ﬁnd a network of any size that runs only one IP routing protocol. When multiple protocols are running, it is necessary to redistribute.
Although the concept of redistribution is straightforward, the design and implementation are extremely tricky. Without a full documented understanding of both the network and the trafﬁc ﬂow, the implementation of redistribution can result in routing loops or the selection of suboptimal paths.
The problems that can occur from redistribution are typically difﬁcult to troubleshoot because the symptoms often appear some distance from the conﬁguration error. The problems experienced as a result of multiple routing processes and their redistribution include the following:
■ The wrong, or less efﬁcient, routing decision is made because of the difference in routing metrics. The choice of the less efﬁcient route is referred to as choosing the suboptimal path.
■ A routing loop occurs, in which the data trafﬁc travels in a circle without ever arriving at the destination. This is normally due to routing feedback, where routers send routing information received from one autonomous system back into the same autonomous system.
■ The convergence time of the network increases because of the different technologies involved. If the routing protocols converge at different rates, this might result in timeouts and the temporary loss of networks.
■ The decision-making process and the information sent within the protocols might be incompatible and not easily exchanged, leading to errors and complex conﬁguration.