Cisco CCNA 2 Exploration - Routing

Transcription

1 Cisco CCNA 2 Exploration - Routing Chapter 9 EIGRP João José jjose@ualg.pt Based on: Graziani, R. (2008) CIS 82 Routing Theory and Concepts

2 Introduction to EIGRP EIGRP: An Enhanced Distance Vector Routing Protocol Autonomous Systems EIGRP Message Format Protocol-Dependent Modules RTP and EIGRP Packet Types Hello Protocol EIGRP Bounded Updates DUAL: An Introduction Administrative Distance Authentication

3 Introduction to EIGRP Enhanced Interior Gateway Routing Protocol (EIGRP) Distance vector Classless routing protocol Released in 1992 with Cisco IOS Software Release Enhancement of Cisco Interior Gateway Routing Protocol (IGRP). Both are Cisco proprietary Operate only on Cisco routers.

4 Introduction to EIGRP The term hybrid routing protocol is sometimes used to define EIGRP. Misleading, not a hybrid between distance vector and link-state Solely a distance vector routing protocol.

5 Introduction to EIGRP Instead of hop count, both IGRP and EIGRP use metrics composed of bandwidth, delay, reliability, and load. Only bandwidth and delay are used by default.

6 Autonomous Systems and Process IDs An autonomous system (AS) is a collection of networks under the administrative control of a single entity that presents a common routing policy to the Internet. Described in RFC AS numbers are assigned by IANA and its RIR. Same authority that assigns IP address space.

7 Autonomous Systems and Process IDs Who needs an autonomous system number? ISPs Internet backbone providers Large institutions connecting to other entities that also have an autonomous system number. Uses exterior gateway routing protocol BGP. The vast majority of companies and institutions with IP networks do not need an autonomous system number because they come under the control of a larger entity such as an ISP. BGP

8 EIGRP Message Format EIGRP Header field Data field = Type/Length/Value, or TLV. Encapsulated in an IP packet. Protocol field = 88 (EIGRP) Destination IP address = multicast If the EIGRP packet is encapsulated in an Ethernet frame: Destination MAC, multicast address: E A

9 EIGRP Packet Header Message Format Note: All fields are shown to provide an accurate picture of the EIGRP message format. However, only the fields relevant to the CCNA candidate are discussed. Opcode specifies the EIGRP packet type as one of the following: 1- Update 3- Query 4- Reply 5- Hello

10 EIGRP Packet Header Message Format Note: All fields are shown to provide an accurate picture of the EIGRP message format. However, only the fields relevant to the CCNA candidate are discussed. Autonomous system number Specifies the EIGRP routing process. Unlike RIP, Cisco routers can run multiple instances of EIGRP. (more later) EIGRP packet types are discussed later in this chapter.

11 EIGRP TLV Message Format EIGRP uses weights for its composite metric. Default, only bandwidth (K1) and delay (K3) are weighted (used). Set to 1. Other K values are set to 0 (affect load and reliability). More later. The hold time Amount of time the EIGRP neighbor receiving this message should wait before considering the advertising router to be down. More later

12 IP Internal Routes TLV Metric fields: Delay and Bandwidth Reliability and Load (more later) Subnet mask field (Prefix Length): Example, the prefix length for is 24 (/24) Destination field: The destination network.

13 IP External Routes TLV In this chapter, we import or redistribute a default static route into EIGRP. Additional fields All the fields used by the IP Internal TLV

14 Note on MTU Some EIGRP literature might incorrectly state that the maximum transmission unit (MTU) is one of the metrics used by EIGRP. MTU is not a metric used by EIGRP. The MTU is included in the routing updates, but it is not used to determine the routing metric.

15 Protocol-Dependent Modules EIGRP uses protocol-dependent modules (PDM). to route different protocols, including: IP, Internetwork Packet Exchange (IPX), AppleTalk PDMs are responsible for the specific routing tasks for each network layer protocol. Example The IP-EIGRP module is responsible for: Sending and receiving EIGRP packets that are encapsulated in IP. Using DUAL to build and maintain the IP routing table.

16 RTP and EIGRP Packet Types Reliable Transport Protocol (RTP) Delivery and reception of EIGRP packets. Cannot use the services of UDP or TCP IPX and AppleTalk do not use protocols from the TCP/IP protocol suite. RTP includes both reliable delivery and unreliable delivery of EIGRP packets: Reliable RTP requires an acknowledgment (like TCP). Unreliable RTP does not require an acknowledgment (like UDP). RTP can send packets either as a unicast or a multicast ( ).

17 EIGRP Packet Types Hello Packet Hello packets are used by EIGRP to: Discover neighbors Form adjacencies with those neighbors EIGRP hello packets: multicasts Unreliable RTP delivery

18 EIGRP Packet Types Acknowledgement Packet Update Query Reply X Update Query Reply Acknowledgment (ACK) Packet Sent back when reliable delivery is used. Reliable delivery: Update, Query, Reply Unreliable delivery: Hello (no acknowledgment packet) Sent as an unreliable unicast.

19 EIGRP Packet Types Update Packet X EIGRP uses triggered updates Update Packets Contains only the routing information needed (a change occurs) Sent only to those routers that require it. Uses reliable RTP delivery. Multicast when sent to multiple routers Unicast when sent to a single router

20 EIGRP Packet Types Query and Reply Packets X Why Query? Another router could be attached to the same LAN. Acknowledgement packets omited, to keep example simple Used by DUAL when searching for networks and other tasks. Queries and replies use reliable delivery. To keep this example simple, acknowledgments were omitted in the graphic. All neighbors must send a reply regardless of whether they have or not a route to the downed network. Queries can use multicast or unicast Replies are always sent as unicast Queries and replies packets are discussed in more detail in CCNP.

21 Hello Protocol Before any EIGRP packets can be exchanged between routers, EIGRP must first discover its neighbors. EIGRP routers discover neighbors and establish adjacencies with neighbor routers using the hello packet.

22 Hello Protocol Most networks, EIGRP hello packets are sent every 5 seconds. On multipoint nonbroadcast multiaccess (NBMA) networks such as X.25, Frame Relay, and ATM interfaces with access links of T1 (1.544 Mbps) or slower, hellos are unicast every 60 seconds. An EIGRP router assumes that as long as it is receiving hello packets from a neighbor, the neighbor and its routes remain viable.

23 Hello Protocol Hold time - maximum time the router should wait to receive the next hello before declaring that neighbor as unreachable. Default hold time - 3 times the hello interval, 15 seconds on most networks 180 seconds on low-speed NBMA networks If the hold time expires: EIGRP declares the route as down DUAL searches for a new path in the topology table or by sending out queries. More later.

24 EIGRP Bounded Updates EIGRP uses the terms partial and bounded when referring to its update packets. EIGRP sends its updates only when the metric for a route changes. The term partial means that the update only includes information about the route changes. The term bounded refers to the propagation of partial updates sent only to those routers that are affected by the change. This minimizes the bandwidth required to send EIGRP packets.

25 DUAL: An Introduction J. J. Garcia-Luna-Aceves Diffusing Update Algorithm (DUAL) is the convergence algorithm used by EIGRP. First proposed by E. W. Dijkstra and C. S. Scholten. The most prominent work with DUAL has been done by J. J. Garcia-Luna- Aceves. Routing loops, even temporary ones, can be extremely detrimental to network performance. Distance vector routing protocols such as RIP prevent routing loops with hold-down timers and split horizon. Although EIGRP uses both of these techniques, it uses them somewhat differently; the primary way that EIGRP prevents routing loops is with the DUAL algorithm.

26 DUAL: An Introduction Example X A directly connected network on R2 goes down. 1. R2 sends an EIGRP update message to its neighbors indicating the network is down. 2. R1 and R3 return an EIGRP acknowledgment indicating that they have received the update from R2. X

27 DUAL: An Introduction Example X 3. R2 doesn t have an EIGRP backup route known as a feasible successor, so R2 sends an EIGRP query to its neighbors asking them whether they have a route to this downed network. 4. R1 and R3 return an EIGRP acknowledgment indicating that they have received the query from R2. 5. R1 and R3 send an EIGRP reply message in response to the query sent by R2. In this case, the reply would state that the router does not have a route to this network. 6. R2 returns an acknowledgment indicating that it received the reply. X

28 Administrative Distance When compared to other interior gateway protocols (IGP), EIGRP is the most preferred by the Cisco IOS (it has the lowest Aministrative Distance). Later in this chapter, you learn how to configure EIGRP summary routes.

29 Authentication Like other routing protocols, EIGRP can be configured for authentication. It is good practice to authenticate transmitted routing information. This practice ensures that routers will accept routing information only from other routers that have been configured with the same password or authentication information. When authentication is configured on a router, the router authenticates the source of each routing update packet that it receives. However, authentication does not encrypt the router s routing table.

30 EIGRP Metric Calculation EIGRP Composite Metric and the K Values EIGRP Metrics Using the bandwidth Command Calculating the EIGRP Metric

31 EIGRP Composite Metric and the K Values EIGRP uses the following values in its composite metric to calculate the preferred path to a network: Bandwidth Delay Reliability Load Note: Although MTU is included in the routing table updates, it is not a routing metric used by EIGRP or IGRP.

32 The Composite Metric By default, K1 and K3 are set to 1, and K2, K4, and K5 are set to 0. The result is that only the bandwidth and delay values are used in the computation of the default composite metric. Note: Modifying the metric weights is beyond the scope of this course, but their relevance is important in establishing neighbors and is discussed in a later section. The tos (type of service) value is left over from IGRP and was never implemented. (The tos value is always set to 0.)

33 Verifying the K Values R1# show ip protocols Routing Protocol is eigrp 1 Outgoing update filter list for all interfaces is not set Incoming update filter list for all interfaces is not set Default networks flagged in outgoing updates Default networks accepted from incoming updates EIGRP metric weight K1=1, K2=0, K3=1, K4=0, K5=0 <output omitted> K1 K2 K3 K4 K5 The K values on R1 are set to the default. Changing these values to other than the default is not recommended unless the network administrator has a very good reason to do so. Cisco recommends that these values are not modified.

34 Examining the Metric Values R1# show interface serial 0/0/0 Serial0/0/0 is up, line protocol is up Hardware is GT96K Serial Description: Link to R2 Internet address is /30 MTU 1500 bytes, BW 1544 Kbit, DLY usec, reliability 255/255, txload 1/255, rxload 1/255 <output omitted> show interface command, lets you can examine the actual values used for bandwidth, delay, reliability, and load in the computation of the routing metric. Default values: bandwidth delay

35 Bandwidth R1# show interface serial 0/0/0 <output omitted> MTU 1500 bytes, BW 1544 Kbit, DLY usec, reliability 255/255, txload 1/255, rxload 1/255 <output omitted> The bandwidth metric (1544 Kbps) is a static value used by some routing protocols such as EIGRP and OSPF to calculate their routing metric. kilobits per second (Kbps). Most serial interfaces use the default bandwidth value of 1544 Kbps or 1,544,000 bps (1.544 Mbps). The value of the bandwidth might or might not reflect the actual physical bandwidth of the interface. Modifying the bandwidth value does not change the actual bandwidth of the link. Should reflect actual bandwidth of the link. (next).

36 Delay R1# show interface serial 0/0/0 <output omitted> MTU 1500 bytes, BW 1544 Kbit, DLY usec, reliability 255/255, txload 1/255, rxload 1/255 <output omitted> Delay is a measure of the time it takes for a packet to traverse a route. Based on the type of link, he interface Expressed in microseconds (10-6 seconds). The router does not actually track how long packets are taking to reach the destination. Like the bandwidth value, delay is a default value that can be changed by the network administrator.

37 Delay R1# show interface serial 0/0/0 <output omitted> MTU 1500 bytes, BW 1544 Kbit, DLY usec, reliability 255/255, txload 1/255, rxload 1/255 <output omitted> 100 microseconds for Fast Ethernet interfaces. Default value is 20,000 microseconds for serial interfaces

38 Reliability R1# show interface serial 0/0/0 <output omitted> MTU 1500 bytes, BW 1544 Kbit, DLY usec, reliability 255/255, txload 1/255, rxload 1/255 <output omitted> Reliability is a measure of the probability that the link will fail or how often the link has experienced errors. Measured dynamically with a value between 0 and 255, with 1 being a minimally reliable link and 255 being 100 percent reliable. Reliability is calculated on a 5-minute weighted average to avoid the sudden impact of high (or low) error rates. Reliability is expressed as a fraction of 255; the higher the value, the more reliable the link. So, 255/255 would be 100 percent reliable, whereas a link of 234/255 would be 91.8 percent reliable. Remember that by default EIGRP does not use reliability in its metric calculation.

39 Load R1# show interface serial 0/0/0 <output omitted> MTU 1500 bytes, BW 1544 Kbit, DLY usec, reliability 255/255, txload 1/255, rxload 1/255 <output omitted> Load reflects the amount of traffic using the link. Value between 0 and 255. Similar to reliability, load is expressed as a fraction of 255. However, in this case, a lower load value is more desirable because it indicates less load on the link. So, 1/255 would be a minimally loaded link. 40/255 is a link at 16 percent capacity, and 255/255 is a link that is 100 percent saturated. Load is displayed as both an outbound, or transmit, load value (txload) and an inbound, or receive, load value (rxload). This value is calculated on a 5-minute weighted average to avoid the sudden impact of high (or low) channel usage. Remember that by default EIGRP does not use load in its metric calculation.

40 Using the bandwidth Command Router(config-if)# bandwidth kilobits Most serial links, the bandwidth metric defaults to 1544 Kbps. Correct value for bandwidth is very important to the accuracy of routing information Use the interface command bandwidth to modify the bandwidth metric. Use the interface command no bandwidth to restore the default value.

41 Using the bandwidth Command R1(config)# inter s 0/0/0 R1(config-if)# bandwidth 64 R2(config)# inter s 0/0/0 R2(config-if)# bandwidth 64 R2(config)# inter s 0/0/1 R2(config-if)# bandwidth 1024 R3(config)# inter s 0/0/1 R3(config-if)# bandwidth 1024 Modify the bandwidth on the appropriate serial interfaces. Be sure to modify both ends of the link.

42 Verify changes R2# show interface serial 0/0/0 Serial0/0/0 is up, line protocol is up Hardware is PowerQUICC Serial Internet address is /30 MTU 1500 bytes, BW 64 Kbit, DLY usec, reliability 255/255, txload 1/255, rxload 1/255 <some output omitted> R2# show interface serial 0/0/1 Serial0/0/1 is up, line protocol is up Hardware is PowerQUICC Serial Internet address is /30 MTU 1500 bytes, BW 1024 Kbit, DLY usec, reliability 255/255, txload 1/255, rxload 1/255 <some output omitted> Verify the change using the show interface command.

43 Calculating the EIGRP Metric

44 Calculating the EIGRP Metric Step 1. Determine the link with the slowest bandwidth on the way to the destination. Step 2. Divide 10,000,000 by the bandwidth, and multiply by 256. (10,000,000/bandwidth) * 256 = compositebandwidth Step 3. Determine the delay value for each outgoing interface on the way to the destination. Step 4. Sum those delay values and divide by 10, then multiply by 256. ( delay/10)* 256 = compositedelay Step 5. Add the portions of the composite metric. So, add compositebandwidth and compositedelay values to obtain the EIGRP metric.

45 Slowest bandwidth Step 1. Determine the link with the slowest bandwidth. The serial 0/0/1 interface on R2 has a bandwidth of 1024 Kbps, or 1,024,000 bps. (Slowest) The Fast Ethernet 0/0 interface on R3 has a bandwidth of 100,000 Kbps, or 100 Mbps.

46 Bandwidth (10,000,000/bandwidth) * 256 = 2,499,840 Slowest bandwidth Step 2. Divide 10,000,000 by the bandwidth, and multiply by 256. Slower bandwidth = 1024 Kbps (10,000,000/bandwidth) * 256 = 2,499,840 (compositebandwidth) The bandwidth portion of the composite metric is 2,499,840

47 Step 3. Determine the delay value for each outgoing interface. The serial 0/0/1 interface on R2 has a delay of 20,000 microseconds. The Fast Ethernet 0/0 interface on R3 has a delay of 100 microseconds.

48 Delay (20,000/ /10)256 = 514,560 Step 3. Sum the delay values and divide by 10, then multiply by 256. serial 0/0/1 R2-20,000, Fast Ethernet 0/0 R3 100 (20, )/10*256 = 514,560 (compositedelay) (( delay/10) * 256) The delay portion of the composite metric is 514,560

49 Route Metric 2,499, ,560 = 3,014,400 Slowest bandwidth Step 4. Add portions of the composite metric: (compositebandwidth+compositedelay) 2,499, ,560, to obtain the EIGRP metric of 3,014,400.

50 Calculating the EIGRP Metric R2# show ip route <code output omitted> D /24 [90/ ] via , 00:00:15, S0/0/1

51 DUAL DUAL Concepts Successor and Feasible Distance Feasible Successors, Feasibility Condition, and Reported Distance Topology Table: Successor and Feasible Successor Topology Table: No Feasible Successor Finite State Machine

52 DUAL Concepts Diffusing Update Algorithm is the algorithm used by EIGRP. Determines: best loop-free path loop-free backup paths (which can be used immediately) DUAL also provides the following: Fast convergence Minimum bandwidth usage with bounded updates DUAL uses several terms that are discussed in more detail throughout this section: Successor Feasible distance Feasible successor Reported distance or advertised distance Feasible condition or feasibility condition

53 Successor and Feasible Distance R2# show ip route <code output omitted> IP address of the successor D /24 [90/ ] via , 00:00:15, S0/0/1 A successor is a neighboring router that is used for packet forwarding and is the least-cost route to the destination network. The IP address of a successor is shown in a routing table entry right after the word via.

54 Successor and Feasible Distance R2# show ip route <code output omitted> Feasible Distance D /24 [90/ ] via , 00:00:15, S0/0/1 Feasible distance (FD) is the lowest calculated metric to reach the destination network. FD is the metric listed in the routing table entry as the second number inside the brackets. As with other routing protocols, this is also known as the metric for the route.

55 Feasible Successors, Feasibility Condition, and Reported Distance DUAL fast convergence because it can use backup paths to other routers known as feasible successors. Does not require to recompute DUAL.

56 Feasible Successors, Feasibility Condition, and Reported Distance Is R1 a Feasible Successor? Does R2 know if R1 has a loop-free backup path to /24? Remember, EIGRP is a Distance Vector Routing protocol.

57 Feasible Successor? A feasible successor (FS) is a neighbor who has a loop-free backup path to the same network as the successor by satisfying the feasibility condition. Would R2 consider R1 to be a feasible successor to network /24? To be a feasible successor, R1 must satisfy the feasibility condition (FC).

58 Feasibility Condition 2. R1: I will send R2 my RD to /24 of 2,172, R1: My FD to /24 is 2,172,416 (my routing table). R1# show ip route <output omitted> D /24 [90/ ] via , 01:12:26, Serial0/0/1 Feasibility Condition: The FC is met when a neighbor s reported distance (RD) to a network is less than the local router s FD to the same destination network. The reported distance (advertised distance) - EIGRP neighbor s FD to the same destination network. The metric that a router reports to a neighbor about its own cost to that network.

59 Feasibility Condition R2: If my path to R3 fails can I reach /24 via R1? Does it meet the FC? If R3 is the successor, can the neighbor R1 be a Feasibile Successor to this same /24 network? In other words, if the link between R2 and R3 fails, can R1 immediately be used as a backup path without a recomputation of DUAL? R1 can only be an FS (Feasible Successor) if it meets the FC (Feasibility Condition).

60 Feasibility Condition 2. R1: I will send R2 my RD to /24 of 2,172, R1: My FD to /24 is 2,172,416 (my routing table). R1# show ip route <output omitted> D /24 [90/ ] via , 01:12:26, Serial0/0/1 R1 FD to /24 is 2,172,416. R1 reports 2,172,416 to R2 as its RD (Reported Distance) R2 s perspective: 2,172,416 is R1 s RD (Reported Distance). R1 s perspective: 2,172,416 is its FD (Feasible Distance).

61 R2# show ip route <output omitted> D /24 [90/ ] via , 00:00:15, Serial0/0/1 R1 s RD is less than R2 s FD, so R1 is the FS for /24 R1# show ip route <output omitted> D /24 [90/ ] via , 01:12:26, Serial0/0/1 Because the RD of R1, 2,172,416, is less than R2 s own FD of 3,014,400, R1 meets the feasibility condition. R1 is now an FS (Feasible Successor) for R2 to the /24 network.

62 R2# show ip route <output omitted> D /24 [90/ ] via , 00:00:15, Serial0/0/1 If R1 was the successor, the 64 Kbps link is the slowest bandwidth, plus another delay of R1 s s0/0/1. This would be a higher metric (41, ) than 3,014,400 via successor R3. R1# show ip route <output omitted> D /24 [90/ ] via , 01:12:26, Serial0/0/1 Why isn t R1 the successor if its RD is less than R2 s FD to /24? The 64 Kbps link would be used as the slowest bandwidth part of the metric calculation. The total cost for R2, its FD to reach /24 is greater through R1 41, than it is through R3 3,014,400.

63 Topology Table: Successor and Feasible Successor R2# show ip eigrp topology IP-EIGRP Topology Table for AS(1)/ID( ) Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply, r - reply Status, s - sia Status <output omitted> P /24, 1 successors, FD is Successor via ( /28160), Serial0/0/1 Feasible Successor via ( / ), Serial0/0/0 P /30, 1 successors, FD is via Connected, Serial0/1 <output omitted> The successor, FD, and any FSs with their RDs are kept by a router in its EIGRP topology table or topology database.

64 Topology Table: Successor and Feasible Successor P: This route is in the passive state. DUAL is not performing its diffusing computations to determine a path for a network The route is in a stable mode All routes should be in this state for stable routing domain. active state - DUAL is recalculating or searching for a new path, /24: This is the destination network that is also found in the routing table. 1 successors: This shows the number of successors for this network. If there are multiple equal-cost paths to this network, there will be multiple successors. FD is : This is the FD, the EIGRP metric to reach the destination network.

65 Successor Information via : This is the next-hop address of the successor, R3. This address is shown in the routing table. 3,014,400: This is the FD to /24. It is the metric shown in the routing table. 28,160: This is the RD of the successor R3 s cost to reach this network. Serial0/0/1: This is the outbound interface used to reach this network. Also shown in the routing table.

66 Feasible Successor Information via : This is the next-hop address of the FS, R1. 41,026,560: This would be R2 s new FD to /24 if R1 became the new successor. 2,172,416: This is the RD of the FS or R1 s metric to reach this network. This value, RD, must be less than the current FD of 3,014,400 to meet the FC. Serial0/0/0: This is the outbound interface used to reach the FC, if this router becomes the successor.

67 Show eigrp topology [network] R2# show ip eigrp topology IP-EIGRP topology entry for /24 State is Passive, Query origin flag is 1, 1 Successor(s), FD is Routing Descriptor Blocks: (Serial0/0/1), from , Send flag is 0x0 Composite metric is ( /28160), Route is Internal Vector metric: Minimum bandwidth is 1024 Kbit Total delay is microseconds Reliability is 255/255 Load is 1/255 Minimum MTU is 1500 Hop count is (Serial0/0/0), from , Send flag is 0x0 Composite metric is ( / ), Route is Internal Vector metric: Minimum bandwidth is 64 Kbit Total delay is microseconds Reliability is 255/255 Load is 1/255 Minimum MTU is 1500 Hop count is 2 Optional metrics Default metrics calculations Successor Other information passed in routing update (not part of composite metric) Feasible Successor Reported Distance of R1 to R2 for Feasible Distance for to if R3 was the successor. This is the metric that would be if R2 s routing table if R3 was the successor.

68 Topology Table: No Feasible Successor R1# show ip route D C D D C D C D /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:45:09, Null /30 is directly connected, Serial0/0/ /30 [90/ ] via , 00:44:56, S0/0/ /16 is variably subnetted, 4 subnets, 3 masks /16 is a summary, 00:46:10, Null /24 is directly connected, FastEthernet0/ /24 [90/ ] via , 00:45:09, S0/0/ /30 is directly connected, Serial0/0/ /24 [90/ ] via , 00:44:55, S0/0/1 Route to /24: Successor is R3 via FD of 2,172,416

69 Topology Table: No Feasible Successor? R1# show ip route D /24 [90/ ] via , 00:44:55, S0/0/1 Is R2 a Feasible Successor? If current path to /24 fails will R1 (DUAL) automatically use R2 as a backup route? Is the path to /24 via R2 a guaranteed loop-free path? (more later) Does R2 meet the Feasibility Condition?

70 Topology Table: No Feasible Successor R1# show ip eigrp topology <output omitted> P /24, 1 successors, FD is <output omitted> via ( /28160), Serial0/0/1 The topology table only shows the successor No Feasible Successors. R2 is not an FS because it does not meet the FC. Topology shows that R2 has a backup route EIGRP does not have a map of the network topology. EIGRP is a distance vector routing protocol

71 Topology Table: No Feasible Successor R1# show ip eigrp topology all-links <output omitted> P /24, 1 successors, FD is , serno 5 Successor Feasible Successor? (No) <output omitted> You can view all possible links whether they satisfy the feasible condition or not by adding the [all-links] option. Even those routes that are not FSs. Is R2 a Feasible Successor? via ( /28160), Serial0/0/1 via ( / ), Serial0/0/0 Does R2 meet the Feasibility Condition? Is R2 s RD less than R1 s FD? R1 s FD R2 s RD

72 Topology Table: No Feasible Successor? R2 s FD is 3,014,400 R2 does not meet FC! 3,014,400 > 2,172,416 R2 s RD to R1 is 3,014,400 R1 s current FD is 2,172,416 R1# show ip eigrp topology all-links <output omitted> P /24, 1 successors, FD is , serno 5 <output omitted> R1 s FD via ( /28160), Serial0/0/1 via ( / ), Serial0/0/0 R2 s RD Even though R2 looks like a viable backup path to /24, R1 has no idea that its path is not a potential loop back through itself.

73 Topology Table: No Feasible Successor? R2 s FD is 3,014,400 R2 does not meet FC! 3,014,400 > 2,172,416 R2 s RD to R1 is 3,014,400 R1 s current FD is 2,172,416 R1# show ip eigrp topology all-links <output omitted> P /24, 1 successors, FD is , serno 5 <output omitted> via ( /28160), Serial0/0/1 via ( / ), Serial0/0/0 Does this mean R2 cannot be used if the successor fails? R3 can be used, but there will be a longer delay before adding it to the routing table. Before this can happen, DUAL will need to do some further processing, which is explained in the next topic.

74 Finite State Machine The centerpiece of EIGRP is DUAL (EIGRP route-calculation engine). DUAL Finite State Machine (FSM) This FSM contains all the logic used to calculate and compare routes in an EIGRP network.

75 DUAL FSM FSMs defines: The set of possible states that something can go through What events cause those states What events result from those states Beyond the scope of this course.

76 FSM Example NO FOOD FOR 5 HOURS GET FOOD HUNGRY (START) FOOD IS INEDIBLE EAT MORE FOOD FULL EATING NOT ENOUGH FOOD EAT ENOUGH FOOD

77 DUAL FSM R2# show ip eigrp topology <partial output> P /24, 1 successors, FD is Successor via ( /28160), Serial0/0/1 Feasible Successor via ( / ), Serial0/0/0 R2 is currently using R3 as the successor to /24. R2 currently lists R1 as an FS

78 DUAL FSM X R2# debug eigrp fsm R2# conf t R2(config)# int s0/0/1 R2(config-if)# shutdown <some debug output omitted> DUAL: Find FS for dest /24. FD is , RD is DUAL: metric / DUAL: metric / found Dmin is DUAL: Removing dest /24, nexthop DUAL: RT installed /24 via shutdown command simulates a failure of the link between R2 and R3.

79 DUAL FSM X R2# debug eigrp fsm R2# conf t R2(config)# int s0/0/1 R2(config-if)# shutdown <some debug output omitted> DUAL: Find FS for dest /24. FD is , RD is DUAL: metric / DUAL: metric / found Dmin is DUAL: Removing dest /24, nexthop DUAL: RT installed /24 via DUAL FSM searches for and finds an FS for the route in the EIGRP topology table. The FS, R1, now becomes the successor and is installed in the routing table as the new best path to /24.

80 DUAL FSM Previous topology table X R2# show ip route <some output omitted> D /24 [90/ ] via , 00:08:58, Serial0/0 R2# show ip eigrp topology <partial output> P /24, 1 successors, FD is Successor via ( / ), Serial0/0 R1 now becomes the successor This route is installed in the routing table as the new best path to /24. There are no new feasible successors.

81 No Feasible Successor R2 s RD to R1 is 3,014,400? R2 s FD is 3,014,400 R1 s current FD is 2,172,416 What if the path to the successor fails and there are no FSs? R1 to /24 R3 is the Successor No Feasible Successors R2 is not a FS does not meet FC. Means that DUAL does not have a guaranteed loop-free backup path to the network, so it wasn t added to the topology table as an FS.

82 No Feasible Successor R2 s RD to R1 is 3,014,400 X? R2 s FD is 3,014,400 R1 s current FD is 2,172,416 R1# show ip eigrp topology <output omitted> P /24, 1 successors, FD is via ( /28160), Serial0/0/1 <output omitted> Simulate a failure of the link between R1 and R3 with a shutdown on R1 s S0/0/1 interface.

83 No Feasible Successor R1# debug eigrp fsm R1# conf t R1(config)# int s0/0/1 R1(config-if)# shutdown <some debug output omitted> DUAL: Find FS for dest /24. FD is , RD is DUAL: metric / DUAL: metric / not found Dmin is DUAL: Dest /24 entering active state. DUAL: rcvreply: /24 via metric / DUAL: Find FS for dest /24. FD is , RD is found DUAL: Removing dest /24, nexthop DUAL: RT installed /24 via /24 network put into the active state and shows that EIGRP queries are sent to other neighbors. R2 replies with a path to this network, This becomes the new successor Installed into the routing table.

84 No Feasible Successor 5. Reply from R2 Yes I have a route to /24 2. Query from R1 for /24 X 3. Query from R2 for /24 4. Reply from R3 Yes I have a route to /24 1. R1 puts /24 into Active State. 5. R1 puts route to /24 via R2 into Routing Table. 1. When the successor is no longer available and there is no FS, DUAL puts the route into active state DUAL will send EIGRP queries asking other routers for a path to this network Other routers return EIGRP replies, letting the sender of the EIGRP query know whether they have a path to the requested network. If none of the EIGRP replies have a path to this network, the sender of the query will not have a route to this network. 5. If the sender of the EIGRP queries receives EIGRP replies that include a path to the requested network, the preferred path is added as the new successor and added to the routing table.

85 No Feasible Successor 5. Reply from R2 Yes I have a route to /24 2. Query from R1 for /24 X 3. Query from R2 for /24 4. Reply from R3 Yes I have a route to /24 1. R1 puts /24 into Active State. 5. R1 puts route to /24 via R2 into Routing Table. R1# show ip route <some output omitted> D /24 [90/ ] via , 00:00:17, Serial0/0/0 This process takes longer than if DUAL had an FS in its topology table and was able to quickly add the new route to the routing table. Note: DUAL FSM and the process of queries and replies is beyond the scope of this course.

86 No Feasible Successor X R1# show ip eigrp topology <parital output> P /24, 1 successors, FD is Successor via ( / ), Serial0/0/0 The topology table for R1 now shows R2 as the successor and shows that there are no new feasible successors.

87 Please check for EIGRP Configuration after the class (last slides)

88 Cisco curriculum available at: (Internet Explorer recommended) After login, under: Course Materials This presentation is available at: Original presentations from:

89 EIGRP Configuration

90 Basic EIGRP Configuration EIGRP Network Topology Autonomous Systems and Process IDs The router eigrp Command The network Command Verifying EIGRP Examining the Routing Table

91 Topology Includes the addition of the ISP router. R1 and R2 routers have subnets that are part of the /16.

92 R1 s running-config hostname R1! interface FastEthernet0/0 ip address ! interface Serial0/0/0 ip address clock rate 64000! interface Serial0/0/1 ip address

93 R2 s running-config hostname R2! interface Loopback1! ip address description Simulated ISP interface FastEthernet0/0! ip address interface Serial0/0/0! ip address interface Serial0/0/1 ip address clockrate ISP router does not physically exist in our configurations. The connection between R2 and ISP is represented with a loopback interface on Router R2.

94 R3 s running-config hostname R3! interface FastEthernet0/0 ip address ! interface Serial0/0/0 ip address clockrate 64000! interface Serial0/0/1 ip address

95 Autonomous Systems and Process IDs An autonomous system (AS) is a collection of networks under the administrative control of a single entity that presents a common routing policy to the Internet. Described in RFC AS numbers are assigned by IANA and its RIR. Same authority that assigns IP address space.

96 Autonomous Systems and Process IDs Who needs an autonomous system number? ISPs Internet backbone providers Large institutions connecting to other entities that also have an autonomous system number. Uses exterior gateway routing protocol BGP. The vast majority of companies and institutions with IP networks do not need an autonomous system number because they come under the control of a larger entity such as an ISP. BGP

97 Process ID Router(config)# router eigrp autonomous-system Router(config)# router eigrp 1 Must be same on all routers in EIGRP routing domain Both EIGRP and OSPF use a process ID to represent an instance of their respective routing protocol running on the router. Although EIGRP refers to the parameter as an autonomous-system number, it actually functions as a process ID. AS parameter is between 1 and 65,535. All routers in this EIGRP routing domain must use the same process ID number (autonomous system number).

98 The router eigrp Command R1(config)# router eigrp 1 R1(config-router)# R2(config)# router eigrp 1 R2(config-router)# R3(config)# router eigrp 1 R3(config-router)# EIGRP is enabled on all three routers using the process ID of 1.

99 The network Command Router(config-router)# network network-address The network command in EIGRP has the same function as in other IGP routing protocols: Any interface on this router that matches the network address in the network command will be enabled to send and receive EIGRP updates. This network (or subnet) will be included in EIGRP routing updates.

100 The network Command Adjacency R1(config-router)# network R2(config-router)# network %DUAL-5-NBRCHANGE: IP-EIGRP 1: Neighbor (Serial0/0) is up: new adjacency The network-address is the classful network address for this interface includes both /24 and /30 subnets. When EIGRP is configured on R2, DUAL sends a notification message to the console stating that a neighbor relationship with another EIGRP router has been established. This new adjacency happens automatically because both R1 and R2 are using the same EIGRP 1 routing process and both routers are now sending updates on the network.

101 The network Command with a Wildcard Mask Router(config-router)# network network-address [wildcard-mask] Network command When uses classful network address: All interfaces on the router that belong to that classful network address will be enabled for EIGRP. To include only specific interface(s), subnets, to be enabled for EIGRP: Use the wildcard-mask option.

102 The network Command with a Wildcard Mask Subtract the subnet mask Wildcard mask R2(config-router)# network Or R2(config-router)# network Think of a wildcard mask as the inverse of a subnet mask. The inverse of subnet mask is To calculate the inverse of the subnet mask, subtract the subnet mask from Some Cisco IOS software versions also let you just enter the subnet mask. However, Cisco IOS software then converts the command to the wildcard mask format, as can be verified with the show runningconfig

103 The network Command with a Wildcard Mask R2(config-router)# network Or R2(config-router)# network R2# show running-config <some output omitted>! router eigrp 1 network network auto-summary

104 The network Command with a Wildcard Mask R2(config-router)# network The passive-interface command should not be used with EIGRP. When the passive-interface command is configured, EIGRP stops sending hello packets on that interface. Will not form an adjacency Unable to send or receive routing updates.

105 Network configurations R1 router eigrp 1 network network R2 router eigrp 1 network network R3 router eigrp 1 network network

106 Verifying EIGRP EIGRP routers must first establish adjacencies with their neighbors before any updates can be sent or received. show ip eigrp neighbors - view the neighbor table and verify that adjacencies with its neighbors. If a neighbor is not listed: Check the local interfaces to make sure it is activated with the show ip interface brief command. Try pinging the IP address of the neighbor.

107 Verifying EIGRP If the ping is successful and EIGRP still does not see the router as a neighbor, examine the following configurations: Are both routers configured with the same EIGRP process ID? Is the directly connected network included in the EIGRP network statements? Is the passive-interface command inappropriately configured, thus preventing EIGRP hello packets on the interface?

108 R1# show ip protocols Routing Protocol is eigrp 1 Outgoing update filter list for all interfaces is not set Incoming update filter list for all interfaces is not set Default networks flagged in outgoing updates Default networks accepted from incoming updates EIGRP metric weight K1=1, K2=0, K3=1, K4=0, K5=0 EIGRP maximum hopcount 100 EIGRP maximum metric variance 1 Redistributing: eigrp 1 Automatic network summarization is in effect Automatic address summarization: /24 for FastEthernet0/0, Serial0/0/0 Summarizing with metric /16 for Serial0/0/1 Summarizing with metric Maximum path: 4 Routing for Networks: Routing Information Sources: Gateway Distance Last Update (this router) 90 00:03: :02:09 Gateway Distance Last Update :02:12 Some items to make note of. These will be explained later.

109 Examining the Routing Table: R1 R1# show ip route Codes: C - connected, S - static, I - IGRP, R - RIP, M - mobile, D - EIGRP, EX - EIGRP external, O - OSPF, <Output omitted> D C D D C D C D /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:03:50, Null /30 is directly connected, Serial0/0/ /30 [90/ ] via ,00:02:43, S0/0/ /16 is variably subnetted, 4 subnets, 3 masks /16 is a summary, 00:10:52, Null /24 is directly connected, FastEthernet0/ /24 [90/ ] via , 00:10:47, S0/0/ /30 is directly connected, Serial0/0/ /24 [90/ ] via , 00:02:31, S0/0/1 Notice that EIGRP routes are denoted in the routing table with a D, which stands for DUAL.

110 Examining the Routing Table: R2 R2# show ip route <Output omitted> D D C D D C C C D /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:04:13, Null /30 [90/ ] via ,00:03:05,S0/0/ /30 is directly connected, Serial0/0/ /16 is variably subnetted, 4 subnets, 3 masks /16 is a summary, 00:04:07, Null /24 [90/ ] via , 00:11:11, S0/0/ /24 is directly connected, FastEthernet0/ /30 is directly connected, Serial0/0/ /30 is subnetted, 1 subnets is directly connected, Loopback /24 [90/ ] via , 00:02:54, S0/0/1 EIGRP is a classless routing protocol (includes the subnet mask in the routing update), it supports variable-length subnet masks (VLSM) and classless interdomain routing (CIDR).

111 Examining the Routing Table: R3 R3# show ip route <Output omitted> D C C D C /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:03:11, Null /30 is directly connected, Serial0/0/ /30 is directly connected, Serial0/0/ /16 [90/ ] via , 00:03:23, S0/0/0 [90/ ] via , 00:03:23, S0/0/ /24 is directly connected, FastEthernet0/0 By default, EIGRP automatically summarizes routes at the major network boundary. You can disable the automatic summarization with the no auto-summary command, just as you can for RIPv2. Null0 summary routes will be explained next.

112 Introducing the Null0 Summary Route R2# show ip route <Output omitted> D D C D D C C C D /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:04:13, Null /30 [90/ ] via ,00:03:05,S0/0/ /30 is directly connected, Serial0/0/ /16 is variably subnetted, 4 subnets, 3 masks /16 is a summary, 00:04:07, Null /24 [90/ ] via , 00:11:11, S0/0/ /24 is directly connected, FastEthernet0/ /30 is directly connected, Serial0/0/ /30 is subnetted, 1 subnets is directly connected, Loopback /24 [90/ ] via , 00:02:54, S0/0/1 The /24 and /16 routes do not actually represent a path to reach the parent networks. If a packet does not match one of the level 2 child routes, it is sent to the Null0 interface. In other words, if the packet matches the level 1 parent, but none of the subnets, the packet is discarded.

113 Introducing the Null0 Summary Route R2# show ip route <Output omitted> D D C D D C C C D /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:04:13, Null /30 [90/ ] via ,00:03:05,S0/0/ /30 is directly connected, Serial0/0/ /16 is variably subnetted, 4 subnets, 3 masks /16 is a summary, 00:04:07, Null /24 [90/ ] via , 00:11:11, S0/0/ /24 is directly connected, FastEthernet0/ /30 is directly connected, Serial0/0/ /30 is subnetted, 1 subnets is directly connected, Loopback /24 [90/ ] via , 00:02:54, S0/0/1 EIGRP automatically includes a Null0 summary route as a child route whenever both of the following conditions exist: There is at least one subnet that was learned via EIGRP. Automatic summarization is enabled. The Null0 summary route is removed when automatic summary is disabled (later).

114 R3 Routing Table R3# show ip route <Output omitted> D C C D C /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:03:11, Null /30 is directly connected, Serial0/0/ /30 is directly connected, Serial0/0/ /16 [90/ ] via , 00:03:23, S0/0/0 [90/ ] via , 00:03:23, S0/0/ /24 is directly connected, FastEthernet0/0 Both R1 and R2 are automatically summarizing the /16 network and sending it as a single routing update. R1 and R2 are not propagating the individual subnets because of automatic summarization. We will turn off automatic summarization later.

115 More EIGRP Configurations The Null0 Summary Route Disabling Automatic Summarization Manual Summarization EIGRP Default Route Fine-Tuning EIGRP

116 The Null0 Summary Route R1# show ip route D C D D C D C /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:45:09, Null /30 is directly connected, Serial0/0/ /30 [90/ ] via , 00:44:56, S0/0/ /16 is variably subnetted, 4 subnets, 3 masks /16 is a summary, 00:46:10, Null /24 is directly connected, FastEthernet0/ /24 [90/ ] via , 00:45:09, S0/0/ /30 is directly connected, Serial0/0/0 D /24 [90/ ] via , 00:44:55, Serial0/0/1 EIGRP automatically includes a Null0 summary route as a child route whenever both of the following conditions exist: There is at least one subnet that was learned via EIGRP. Automatic summarization is enabled. (By default with EIGRP) R1 will discard any packets that match the parent /16 classful network but do not match one of the child routes /24, /24, or /24. For example, a packet to would be discarded.

117 The Null0 Summary Route R1# show ip route D C D D C D C /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 00:45:09, Null /30 is directly connected, Serial0/0/ /30 [90/ ] via , 00:44:56, S0/0/ /16 is variably subnetted, 4 subnets, 3 masks /16 is a summary, 00:46:10, Null /24 is directly connected, FastEthernet0/ /24 [90/ ] via , 00:45:09, S0/0/ /30 is directly connected, Serial0/0/0 D /24 [90/ ] via , 00:44:55, Serial0/0/1 This Null0 summary route is a child route that will match any possible packets of the parent route that do not match another child route. This is regardless of ip classless or no ip classless command. Therefore denying the use of any supernet or default route.

118 Disabling Automatic Summarization /16 R3# show ip route /24 is variably subnetted, 3 subnets, 2 masks D /24 is a summary, 01:08:35, Null0 C /30 is directly connected, Serial0/0/0 C /30 is directly connected, Serial0/0/1 D /16 [90/ ] via , 01:08:30, Serial0/0/0 C /24 is directly connected, FastEthernet0/0 Like RIP, EIGRP automatically summarizes at major network boundaries using the default auto-summary command.

119 Disabling Automatic Summarization /16 R3# show ip route D C C D C / /24 is variably subnetted, 3 subnets, 2 masks /24 is a summary, 01:08:35, Null /30 is directly connected, Serial0/0/ /30 is directly connected, Serial0/0/ /16 [90/ ] via , 01:08:30, Serial0/0/ /24 is directly connected, FastEthernet0/0 Both R1 and R2 automatically summarized those subnets to the /16 classful boundary when sending EIGRP update packets to R3. The result is that R3 has one route to /16 through R1. R1 is the successor because of the difference in bandwidth.

120 Disabling Automatic Summarization /16 R3# show ip route <output omitted> D /16 [90/ ] via , 01:08:30, Serial0/0/0 You can quickly see that this route is not optimal. R3 will route all packets destined for through R1. Across a very slow link to R2 (64 Kbps). Need R1 and R2 to send individual routes for each of the /16 subnets. In other words, R1 and R2 must stop automatically summarizing /16.

121 Disabling Automatic Summarization R1(config)# router eigrp 1 R1(config-router)# no auto-summary %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor (Serial0/0/0) is resync: summary configured %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor (Serial0/0/0) is down: peer restarted %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor (Serial0/0/0) is up: new adjacency <output omitted> R2(config)# router eigrp 1 R2(config-router)# no auto-summary R3(config)# router eigrp 1 R3(config-router)# no auto-summary Automatic summarization can be disabled with the no auto-summary. The router configuration command eigrp log-neighborchanges is on by default on some IOS implementations. If on, you will see output similar to that shown for R1.

122 Disabling Automatic Summarization R1# show ip route /30 is subnetted, 2 subnets C is directly connected, Serial0/0/1 D [90/ ] via , 00:16:55, S0/0/ /16 is variably subnetted, 3 subnets, 2 masks C /24 is directly connected, FastEthernet0/0 D /24 [90/ ] via , 00:16:53, S0/0/1 C /30 is directly connected, Serial0/0/0 D /24 [90/ ] via , 00:16:52, Serial0/0/1 R1 no more Null0 summary routes: D /24 is a summary, 00:45:09, Null0 D /16 is a summary, 00:46:10, Null0 This mean any packets for their parent networks that do not match a child route, the routing table will check supernet and default routes. Unless no ip classess is used

123 Disabling Automatic Summarization R2# show ip route /30 is subnetted, 2 subnets D [90/ ] via , 00:15:44, S0/0/1 C is directly connected, Serial0/0/ /16 is variably subnetted, 3 subnets, 2 masks D /24 [90/ ] via , 00:15:44, S0/0/1 C /24 is directly connected, FastEthernet0/0 C /30 is directly connected, Serial0/0/ /30 is subnetted, 1 subnets C is directly connected, Loopback1 D /24 [90/ ] via , 00:15:44, S0/0/1 R2 no more Null0 summary routes : D /24 is a summary, 00:00:15, Null0 D /16 is a summary, 00:00:15, Null0

124 Disabling Automatic Summarization R3# show ip route C C D D D /30 is subnetted, 2 subnets is directly connected, Serial0/0/ is directly connected, Serial0/0/ /16 is variably subnetted, 3 subnets, 2 masks /24 [90/ ] via , 00:00:11, S0/0/ /24 [90/ ] via , 00:00:12, S0/0/ /30 [90/ ] via , 00:00:12, S0/0/0 [90/ ] via , 00:00:12, S0/0/1 C /24 is directly connected, FastEthernet0/0 Without automatic summarization, R3 s routing table now includes the three subnets: /24, /24, and /24. Why does R3 s routing table now have two equal-cost paths to /24? Shouldn t the best path only be through R1 with the 1544-Mbps link?

125 Disabling Automatic Summarization / /24 R3# show ip route <output omitted> D /30 [90/ ] via , 00:00:12, S0/0/0 [90/ ] via , 00:00:12, S0/0/1 The slowest link is the 64-Kbps link that contains the /30 network. The 1544-Mbps link and the 1024-Kbps link are irrelevant in the calculation as far as the bandwidth metric is concerned.

126 Manual Summarization EIGRP can be configured to summarize routes, whether or not automatic summarization (auto-summary) is enabled. Modified topology.

127 Manual Summarization R3(config)# interface loopback 2 R3(config-if)# ip address R3(config-if)# interface loopback 3 R3(config-if)# ip address R3(config-if)# router eigrp 1 R3(config-router)# network R3(config-router)# network Add two more networks to R3. With the appropriate network commands R3 will propagate these networks to other routers.

128 Manual Summarization /24, /24, / /24, /24, /24 Only pertinent routes shown R1# show ip route D /24 [90/ ] via , 02:07:38, S0/0/1 D /24 [90/ ] via , 00:00:34, S0/0/1 D /24 [90/ ] via , 00:00:18, S0/0/1 R2# show ip route D D D /24 [90/ ] via , 02:08:50, S0/0/ /24 [90/ ] via , 00:01:46, S0/0/ /24 [90/ ] via , 00:01:30, S0/0/1 R1 and R2 routing tables show these additional networks in their routing tables. Instead of sending three separate networks, R3 can summarize the /24, /24, and /24 networks as a single route.

129 Determining the Summary EIGRP Route 1. Write out the networks that you want to summarize in binary. 2. Find the matching bits. Count the number of leftmost matching bits, which in this example is 22. This number becomes your subnet mask for the summarized route: /22 or To find the network address for summarization, copy the matching 22 bits and add all 0 bits to the end to make 32 bits. The result is the summary network address and mask for /22