03 - Enhanced Interior Gateway Routing Protocol

8/3/2019 03 - Enhanced Interior Gateway Routing Protocol

1/64

Enhanced Interior GatewayRouting Protocol

(EIGRP)


2/64

Enhanced Interior Gateway Routing Protocol

Successor to Interior Gateway Routing Protocol (IGRP)

Cisco proprietary hybrid protocol

Both Distance Vector and Link State (Active Adjacency)behaviour

Really Advanced Distance Vector

Classless protocol

Supports VLSM and summarization

What is EIGRP?


3/64

Guarantees loop-free topology

Diffusing Update Algorithm (DUAL)

Fast convergence Fastest of all IGP in certain designs

Reliable & Efficient Updating

Forms active neighbor adjacencies

Guarantees packet delivery with Reliable TransportProtocol (RTP)

Supports partial updates

Not all neighbors need all routes

Why Use EIGRP?


4/64

Multiple routed protocol support

IPv4, IPX, & Appletalk

Legacy now, but originally important in non-convergednetworks

Granular Metric

Hybrid metric derived from multiple factors

(bandwidth -K1, reliability -K2, delay -K3, load -K4, MTU -K5)

Seamless Connectivity across all data link layerprotocols and topologies

No special configuration is required

Why Use EIGRP? (cont.)


5/64

Unequal Cost Load Balancing

Only IGP that supports true load distribution

Control Plane Security

Supports MD5 based authentication

Why Use EIGRP? (cont.)


6/64

EIGRP sends out five different types of packets

Hello (5): Identifies neighbors and serves as a keep-alive

mechanism Update (1): Reliably sending route information

Query (3): Reliably requests specific route information

Reply (4): Reliably responds to a query

Ack (8): Non reliable packets used to acknowledgeupdate, query and replies

In detail

EIGRP Packet types


7/64

Hello packets are used for neighbor discovery andrecovery process

Sent as multicast

Unreliable Do not need acknowledgement packets

5 Seconds on Fast network links

60 seconds on Slow links (< T1 links)

3 times hello times is dead Can be changed using

ip hello-interval eigrp / ip hold-time eigrp interface command

Hello Packets


8/64

Hello packet Capture


9/64

Convey Route information

Sent to communicate routes that a router has used toconverge

Update is sent to affected routers

Updates are sent as multicast when a new route isdiscovered and convergence is completed

Updates are sent as unicast during synchronization oftopology tables at startup sequence

Updates are sent reliably

Update Packet


10/64

Update Packet


11/64

Query packet is generated when performing

Route computation

No Feasible successor Sends query packet to its connected neighbors if they

have a successor to the destination

Queries are multicast, can send unicast in certain

conditions Reliable packet and needs an acknowledgement

Query Packet


12/64

Reply is generated and sent across in response to a

query packet Reply packets are always unicast

Only reply to the originator of the query

Reliable packet, and needs an acknowledgement

Reply Packet


13/64

Used to ACK updates, queries and replies

Unicast hello packets contain a nonzeroacknowledgement number

Hello packet and ACK packet do not requireacknowledgements

A reliable packet is not acknowledged, EIGRPperiodically retransmits packet to neighbor as unicast,

EIGRP window size = 1, no other traffic is sent to thisneighbor

16 unacknowledged retransmissions, neighbor isremoved from neighbor table

Acknowledgement Packet


14/64

EIGRP uses three data table structures

Neighbor table: built from EIGRP Hellos and used for reliabledelivery

show ip eigrp neighbors [listed devices are ready to exchangetopology information]

Topology table: contains EIGRP routing information for thebest paths and loop-free alternatives

Show ip eigrp topology all-links

Show ip eigrp topology

Entry for a prefix can exists in two forms Active / Passive

Routing table: Places best routes from its topology table intothe common routing table

EIGRP Data Table Structures


15/64

Route is passive when the router is not performing

re-computation on that route Route will be active when it is undergoing re-

computation (looking for an alternate path)

Note: Passive state is operational and Stable state

A route will never go into active state if there arefeasible successor

Passive Vs. Active Routes


16/64

Step 1 - Discover EIGRP Neighbors

Step 2 - Exchange Topology Information Step 3 - Choose Best Path via DUAL

Step 4 - Neighbor and Topology Table Maintenance

How EIGRP Works


17/64

EIGRP uses multicast HELLO packets to discover neighbors onEIGRP enabled attached links Transport via IP protocol 88 (EIGRP) Destination address 224.0.0.10

Hello packets contain Autonomous System Number Hold Time (do not have to match) Authentication Metric Weightings (K values)

Neighbors found are inserted into EIGRP neighbor table show ip eigrp neighbors

Neighbors that agree on attributes and exchange updates formactive adjacency

Step 1 -- Discovering EIGRP Neighbors


18/64

Once neighbors are found, EIGRP UPDATE messages used to exchangeroutes Sent as multicast to 224.0.0.10 or as unicast

RTP uses sequence numbers and acknowledgements (ACKs) to ensuredelivery Update messages describe attributes of a route Prefix + Length Next-Hop Bandwidth Delay Load Reliability

MTU Hop Count External Attributes

All routes learned from all neighbors make up the EIGRP topology table show ip eigrp topology

Step 2 -- Exchanging TopologyInformation


19/64

Once topology is learned, DUAL runs to choose loop-free best path toeach destination

Unlike other protocols, EIGRP uses complex composite metric tochoose best path

Composite metric calculated from Administrative Weighting Bandwidth Delay Load Reliability

MTU Path with lowest composite metric is considered best and installed in IP

routing table One or more backup routes can also be pre-calculated per destination Only best route is advertised to other EIGRP neighbors

Step 3 -- Choosing The Best Path


20/64

Unlike RIP or IGRP, active EIGRP neighbor adjacencyreduces convergence time in event of network failure

Adjacent neighborshello packets contain hold time If no hello is received within hold time, neighbor declared

unreachable

When neighbor is lost Paths via that neighbor are removed from topology and

routing table

If backup routes exist, they become new best paths and areinserted in routing table In this case EIGRP can have sub-second convergence

If no backup routes exist, DUAL must run again

Step 4 -- Neighbor and TopologyTable Maintenance


21/64

When best path is lost and no backup routes exist, route goes intoactive state and active timer starts Stable routes not in active state are considered passive

EIGRP QUERY message is reliably sent to remaining neighbors

asking if there is an alternate route QUERY is propagated to all neighbors within EIGRP query

domain or flooding domain More on this later

Neighbors respond with EIGRP REPLY packet indicating ifalternate route is available If alternate route exists, DUAL recalculates new best path If no alternate route, prefix removed from topology table If active timer expires and no REPLY received, route is declared

Stuck-In-Active (SIA) and removed from topology table

DUAL Reconvergence


22/64

EIGRP guarantees loop-free topology through usage

of Split Horizon

Dont advertise routes out the link they came in on

DUAL Feasibility Condition

If your metric is lower than mine, you are loop-free

EIGRP Loop Prevention


23/64

Successor Nearest router to a destination

Feasible Distance (FD) Composite metric of best path

Feasible Successor (FS) Next nearest router to adestination

Advertised Distance (AD) Composite metric learnedfrom neighbor

Local Distance (LD) Composite metric to reach localneighbor

Feasibility Condition (FC) Criteria for valid backuppaths

DUAL Terms


24/64

Local Distance

1520

R1 R3 R5

VLAN X

1


25/64

Local Distance

15LD =20

R1 R3 R5

VLAN X

1

LocalDistance costto reach R3 is

20


26/64

Local Distance

LD=15LD =20

R1 R3 R5

VLAN X

1

Local Distance

cost to reach R1 is20 and R5 is 15


27/64

Local Distance

LD=15LD=20

R1 R3 R5

VLAN X

LD=1

Local Distancecost to reach R3is 15 and VLAN X

is 1


28/64

Advertised Distance

LD=15LD=20

R1 R3 R5

VLAN X

LD=1

Local Distancecost to reach

VLAN X is 1


29/64

Advertised Distance

R1 R3 R5

VLAN X


VLAN X is 1

Advertiseddistance by R5 toreach VLAN X is 1

LD =15LD =20 LD =1AD VLAN X = 1


30/64

Advertised Distance

R1 R3 R5

VLAN X


VLAN X is 1

Advertiseddistance by R5

to reach VLAN Xis 1

LD =15LD =20 LD =1AD VLAN X = 1

Advertiseddistance by R3 toreach VLAN X is

15 (LD), 1(AD)

AD VLAN X = R3sLD + R5s AD


31/64

Feasible Distance

LD =15LD =20

R1 R3 R5

VLAN X

LD =1


VLAN X is 1

AD for VLAN X is 1cost to reach

VLAN X is LD+AD

AD VLAN X = 1

Feasible Distance: Metric of Successor + LD (END to END Cost)


32/64

Feasible Distance

LD =15LD =20

R1 R3 R5

VLAN X

LD =1


VLAN X is 1

AD VLAN X = 1

FeasibleDistance

is AD + LD



33/64

Feasible Distance

LD =15LD =20

R1 R3 R5

VLAN X

LD =1


VLAN X is 1

AD VLAN X = 1

Feasible Distance forVLAN X is

LD to R3 + AD by R3 for

VLAN X



34/64

Feasible Distance

LD =15LD =20

R1 R3 R5

VLAN X

LD =1


VLAN X is 1

AD VLAN X = 1

Feasible Distance forVLAN X is

LD to R3 + AD by R3 for

VLAN X

Note: It is FD that affects the selection of the best routes forincorporation in the routing table. AD is used only to calculate FD


35/64

Neighboring router to reach prefix

The route installed in routing table Least path cost to prefix

Has lowest FD of all possible paths to a prefix

Many entries in topology table with equal cost FD for a

prefix All successors (up to 4) for that destination in routing table

Successor


36/64

A EIGRP router with a back up route to the prefix

Route must be loop free Must not loop to current successor

FS are selected when successors are identified

FS paths are placed in topology table

Topology table retains multiple FS for a prefix

Requirements: Must Qualify Feasibility condition

Feasible Successor


37/64

The Neighborsadvertised distancemust be less thanfeasible distance ofthe current successorfor a prefix

Feasibility condition

10.1.1.0/24

Cost to10.1.1.0/24 is

1000

Cost to10.1.1.0/24

is 1500

LD = 1000

LD = 1000

FD for 10.1.1.0/24via A is 2000

FD for 10.1.1.0/24via B is 2500

A

B


38/64

Feasibility condition

10.1.1.0/24

A is theSuccessor

Cost to10.1.1.0/24

is 1500

LD = 1000

LD = 1000

A

B

Cost to10.1.1.0/24 is

1000

B is the

FeasibleSuccessor


39/64

Once adjacency occurs and update messages are

exchanged, path selection begins Each update includes the metric the upstream router

uses to reach destination (AD)

Local router knows the metric to reach each

upstream router (LD) Best path (successor) is chosen based on lowest AD +

LD

DUAL Path Selection in Detail


40/64

DUAL Example

10

10

15

20

1015 20

20

R1

R2 R3 R4

R5

VLAN x

Local DistanceAdvertised DistanceFeasible Distance

R5 x = 1


41/64

DUAL Example

10

10

15

20

1015 20

20

R1

R2 R3 R4

R5

VLAN x


R5 x = 1

11

1


42/64

DUAL Example

10

10

15

20

1015 20

20

R1

R2 R3 R4

R5

VLAN x


R5 x = 1

11

1

R2R5 x = 11R2R3R5 x = 26

11

11


43/64

DUAL Example

10

10

15

20

1015 20

20

R1

R2 R3 R4

R5

VLAN x


R5 x = 1

11

1

11

11

R2R5 x = 11R2R3R5 x = 26

R3R2R5 x = 21R3R5 x = 16


44/64

DUAL Example

10

10

15

10

1015 20

20R1

R2 R3 R4

R5

VLAN x


R5 x = 1

11

1

11

11

R2R5 x = 11R2R3R5 x = 26

R3R2R5 x = 21R3R5 x = 16R3R4R5x=31

R4R3R5 x = 26R4R5 x = 21

21

21


45/64

DUAL Example

10

10

15

10

1015 20

20R1

R2 R3 R4

R5

VLAN x


R5 x = 1

11

1

11

11

R2R5 x = 11R2R3R5 x = 26

R3R2R5 x = 21R3R5 x = 16R3R4R5x=31

R4R3R5 x = 26R4R5 x = 21

21

2116


46/64

DUAL Example

10

10

15

10

1015 20

20R1

R2 R3 R4

R5

VLAN x


R5 x = 1

11

1

11

11

R2R5 x = 11R2R3R5 x = 26

R3R2R5 x = 21R3R5 x = 16R3R4R5x=31

R4R3R5 x = 26R4R5 x = 21

21

2116

AD for R1:11 vs. 16 vs. 21

FD to x:11+10 vs. 16+20 vs.21+15


47/64

DUAL Example

10

10

15

10

1015 20

20R1

R2 R3 R4

R5

VLAN x


R5 x = 1

11

1

11

11

R2R5 x = 11R2R3R5 x = 26

R3R2R5 x = 21R3R5 x = 16R3R4R5x=31

R4R3R5 x = 26R4R5 x = 21

21

2116

R1

R2

R5

x = 21R1R3R5 x = 36R1R4R5 x = 36

l


48/64

DUAL Example

10

10

15

10

1015 20

20R1

R2 R3 R4

R5

VLAN x


R5 x = 1

FD to x:Via R2 (21) vs. Via R3 (36) vs. R4(36)

Successor


49/64

Once best path is chosen, additional paths areexamined for backup routes

Feasibility Condition (FC) finds loop-free backup

routes via logic If AD < FD, path is loop-free and viable backup

e.g. if your metric is lower than mine, you are closer tothe destination and loop-free

Paths that meet the FC are Feasible Successors (FS) Only Feasible Successors can be used for unequal cost

load balancing

Feasibility Condition in Detail

l


50/64

DUAL Example

10

10

15

10

1015 20

20R1

R2 R3 R4

R5

VLAN x


R5 x = 1

R3R2R5 x = 21R3R5 x = 16R3R4R5x=31

R4R3R5 x = 26R4R5 x = 21

2116

R1

R2

R5

x = 21R1R3R5 x = 36R1R4R5 x = 36

Successor

DUAL E l


51/64

DUAL Example

10

10

15

10

1015 20

20R1

R2 R3 R4

R5

VLAN x


R5 x = 1

R3R2R5 x = 21R3R5 x = 16R3R4R5x=31

R4R3R5 x = 26R4R5 x = 21

2116

R1R2R5 x = 21

R1R3R5 x = 36R1R4R5 x = 36 FD = 21

Find path with AD < 21

Successor

X via R3 = 16

X via R4 = 21

DUAL E l


52/64

DUAL Example

10

10

15

10

1015 20

20R1

R2 R3 R4

R5

VLAN x


R5 x = 1

R4R3R5 x = 26R4R5 x = 21

2116

R1R2R5 x = 21

R1R3R5 x = 36R1R4R5 x = 36 FD = 21

Find path with AD < 21

Successor

X via R3 = 16

X via R4 = 21

FeasibleSuccessor


53/64

Composite Metric - EIGRP

256 1 +2

256 + 3

5

+ 4

bw = 107 / minimum path bandwidth in kbps delay = interface delay in secs / 10

The concept of composite metric is always misunderstood


54/64

Bandwidth: Slowest between Source and Destination

Load: Worst load on a link between Source andDestination

Delay: Cumulative Delay along the path

Reliability: The worst reliability along the path

MTU: Smallest MTU on the path (not used in metriccalculation)

In detail

Metric Parameters

Composite Metric Calculation in Detail


55/64

Unlike other IGPs hop count or BW-based cost, EIGRP metric is a hybrid valuecomprised of

Lowest bandwidth along path in Kbps scaled by 107 * 256

Cumulative delay along path in tens of microseconds (s) scaled by 256

Worst load along path

Worst reliability along path Composite metric is computed as

metric = [k1 * bandwidth + (k2 * bandwidth)/(256 - load) + k3 * delay]

If k5 != 0, metric = metric * [k5/(reliability + k4)]

K values allow for manual administrative weighting

Must match for adjacency to occur Default K values are K1 = 1, K2 = 0, K3 = 1, K4 = 0, K5 = 0

Implies default composite is bandwidth + delay

Reliability and load typically not used since they are constantly changing

Composite Metric Calculation in Detail

d d h


56/64

Bandwidth

Bandwidth: Slowest between Source and Destination

100,000 Kbps

1544 Kbps

R1 R2 R3 R4

107 / BW = 10,000,000 / 1544 = 156250

Bandwidth is taken from configured interface BW

R1#sh int serial 0/0 | in BW

MTU 1500 bytes, BW 1544 Kbit, DLY 20000 usec,

R1#sh int ethernet 0/0 | in BW


R1#sh int fastEthernet 0/0 | in BW


1544 Kbps100,000 Kbps 100,000 Kbps

Link Type Bandwidth Delay

10 Mbps Ethernet 10000 Kpbs 1000

100 Mbps Ethernet 100000 Kpbs 100

1 Gbps Ethernet 1000000 Kbps 10

Serial (


57/64

Worst load on a link between Source and Destination

Huge file transfer, load will vary

Load constantly keeps changing and DUAL calculatesfor successor, feasible successor

Not advised to use in metric calculation

R1#sh int fastEthernet 0/0 | in load

reliability 255/255, txload 1/255, rxload 1/255

Load


58/64

Delay: Cumulative Delay along the path

Represented in 10s microseconds R1 R2 100secs / 10 == 10

R2 R3 20000secs / 10 == 2000

R3 R4 100secs / 10 == 10

10 + 2000 + 10 = 2020

Delay

R1 R2 R3 R4

1544 Kbps100,000 Kbps 100,000 Kbps

100 secs 100 secs20000 secs


59/64

Reliability

The worst reliability along the path

Most of the times it is 255

Depends on keep alive between neighbors

Link flaps will change the reliability of the link

Not advised to use in metric calculation

Reliability will be calculated and informed to upper layerby Physical layer algorithms


60/64

MTU Smallest MTU on the path

maximum size of IP packet that can be transmittedwithout fragmentation

Media Maximum TransmissionUnit (bytes) Notes

Internet IPv4 Path MTU At least 68Practical path MTUs are generally higher. IPv4 links must be able to

forward packets of size up to 68 bytes. Systems may use Path MTU

Discovery to find the actual path MTU. This should not be mistaken

with the packet size every host must be able to handle, which is 576.

Internet IPv6 Path MTU At least 1280 Practical path MTUs are generally higher. Systems must use PathMTU Discovery to find the actual path MTU.

Ethernet v2 1500 Nearly all IP over Ethernet implementations use the Ethernet V2frame format.Ethernet (802.3) 1492

Ethernet Jumbo Frames 1500-9000The limit varies by vendor. For correct interoperation, the whole

Ethernet network must have the same MTU. Jumbo frames are

usually only seen in special purpose networks.WLAN (802.11) 2272

Token Ring (802.5) 4464FDDI 4352

Composite Metric


61/64

All Links are Fast EthernetBW = 100,000 KbpsDLY = 100 Sec

Loopback = 10,000,000Kbps

DLY = 5000 Sec

Advertised DistanceFeasible Distance

Composite Metriccalculation

Loopback R1

R2 R3 R4

R5 Loopback: 5.5.5.5/32

256 1 +2

256 + 3

5

+ 4

Composite Metric


62/64



DLY = 5000 Sec



Loopback R1

R2 R3 R4

R5 Loopback: 5.5.5.5/32

256 1 +2

256 + 3

5

+ 4

R5s Metric Calculation to reach 5.5.5.5/32Values: K1 = K3 = 1 (default)

bw =07

0,000,000=

0,000,000

0,000,000 = 1

delay =000

0= 500 Sec

Substituting bw and delay into the give formula

256 * ( 1*1 + 1* 500) = 256 * ( 1 + 500) = 256 * (501) = 128256

Composite Metric


63/64



DLY = 5000 Sec



Loopback R1

R2 R3 R4

R5 Loopback: 5.5.5.5/32

256 1 +2

256 + 3

5

+ 4

CostforLoopbackis128256(AD)

Composite Metric


64/64



DLY = 5000 Sec



Loopback R1

R2 R3 R4

R5 Loopback: 5.5.5.5/32

2 5

CostforLoopbackis128

256(AD)

R2s Metric Calculation to reach 5.5.5.5/32Values: K1 = K3 = 1 (default)

bw =07

00,000=

0,000,000

00,000 = 100

delay =000

0= 500 Sec

Substituting bw and delay into the give formula

256 * ( 1*100 + 1* 500) = 256 * ( 1 + 500) = 256 * (501) =8 6

Documents

03 - Enhanced Interior Gateway Routing Protocol