EIGRP : Enhanced Interior Gateway Routing Protocol Enhanced interior gateway routing protocol, also called enhanced interior gateway routing protocol. It is an enterprise-level / advanced routing protocol. One of the most powerful IGP (Interior Gateway Protocol) protocols is Cisco's private protocol (it has been publicized in 2013)
EIGRP Features
1.Advanced distance vector
Advanced distance vector type: It is the distance vector type that sends routing information directly. The bottom layer of
Fast convergence
uses Diffusing Update Algorithm, referred to as dual diffusion update algorithm, and the convergence speed is very fast. If the network is set up well, it can be completed within 2s. Convergence, the faster the convergence speed, the more stable the network
3.
Support for VLSM and discontiguous subnets support VLSM and discontinuous subnet (classless routing protocol) advanced routing protocol prerequisites
4. Partial updatesd
triggered, partial, yes Boundary, incremental update (rip is periodic update, shortcoming, periodic update after convergence is complete and stable, wasting network resources), after convergence is completed, only the routing change will occur and the changed route will be sent to the affected router Information to reduce the waste of network resources 5. Support
for multiple network-layer protocols
supports multiple network layer protocols, three-layer protocols, not only IP protocols, but also non-mainstream protocols such as ipx and appletalk
6. Flexible network design
Scalable network Design, strong scalability, there are not many design requirements
7. Multicast and unicast inst ead of broadcast address
uses multicast and unicast instead of broadcast, RIP v2 multicast address: 224.0.0.9, EIGRP multicast address: 224.0.0.10, advanced routing protocols use multicast or unicast
8. Manual summarization at any point
Can support manual network summarization at any point in the network, and support supernet summarization (RIP only supports subnet summarization)
9.100% loop-free classless routing
100 % loop-free classless routing
10.Easy configuration for WANs and LANs
simple WAN and LAN configuration
11. Load balancing across equal and unequal-cost pathways
and other overhead and unequal overhead load balancing, two lines, one is 100M, one 50M, will load balance according to bandwidth
EIGRP neighbor table
two routes exchange routing information with each other, must meet certain conditions, only meet these conditions, the network is stable and reliable.
Before routing and exchanging data, establish an adjacency relationship. Certain conditions must be met to establish an adjacency relationship.
Advanced routing protocols will all have the concept of adjacency relationship. The purpose is to check the condition
. , Check the judgment conditions, and start to synchronize routing information after establishing the adjacency relationship. According to the neighbor table, check the neighbor conditions, and the neighbor is called the next hop.
The EIGRP Topology topology table Table
neighbor table saves directly connected routing information, all the learned routing information is placed in the topology table (there are usable, optimal, sub-optimal, unusable, and even loop routing)
Run the dual algorithm on the topology table to select the optimal loop-free route and put it in the third table route table
Neighbor Neighbor Table and Topology Topology Table are maintained by EIGRP itself, the routing table is global, but the neighbor table and topology table will affect the routing table
DUAL Terminology Terminology
1. Choose the optimal, loop-free path
to the target network
2. AD =cost of the next hop to the target network (Metric)3. FD =cost of thelocal router to the target network = local route to reach the advertised route (Next-hop routing) overhead
4. Lowest-Cost = Lowest-FD
5. (Current) successor = the next-hop router (optimal successor / routing) that
reaches the target network isoptimal6.Feasible successor = the one that reaches the target network Suboptimal next-hop router (feasible successor / backup route, not all networks have backup routes, such as when there is only one feasible route)
7.FC feasible condition: AD <FDmin anti-loop condition
EIGRP Packets
EIGRP packets directly use the IP protocol , protocol number 88
1.Hello
found to maintain the adjacency relationship, is the only periodically sent data packet of EIGRP, the data packet is very small, according to the bandwidth speed of the transmission
bandwidth is greater than T1, every 5s Send once, the bandwidth is less than T1, then send it every 60s, if the data content of the neighbor is not received within 3 times of the hello time, then the interface will be down
2. Upadte
sends routing updates (routing information), when the two routers have just established a connection, when the routing information is synchronized, or when the network topology changes.
3.Query
queries neighbors for specific routing information, and appears in pairs with Reply.
4.Reply
response to Query query
2-4 must be reliable transmission, using RTP protocol (reliable transmission protocol), that is, you must receive an ACK confirmation packet after sending, otherwise continue to resend
5.ACK
confirmation packet
Why EIGRP is called DUAL diffusion update algorithm.
When a route is unreachable and there is no successor, it will perform a diffusion operation. EIGRP must wait for all queries to be returned before returning to the state. It can wait at most 3 minutes and wait 3 minutes All return to the state, it will be converted to SIA (stack in active), this route is not available.
EIGRP Key Techologies
1. Neighbor discovery / recovery
uses Hello packet discovery to maintain adjacency
2. Reliable Transmission Protocol to
ensure the reliability of EIGRP packet transmission
3. DUAL finite-state machine
The minimum cost of thefinite state machineto reach the target network, no loop Path
4. Protocol Dependent Modules (PDMs)
EIGRP not only supports IP, but also supports AppleTalk, Novell NetWare and other unusual protocols.
Each protocol uses independent EIGRP modules to run
EIGRP Metric Calculation
使用复合度量值（bandwidth，relay， reliability，load，MTU）
1.By default ，ERGIP metric：
Metric = bandwidth（slowest link）+ delay（sum of delays）
2.Delay = sum of the delays in the path， in tens of microseconds，multiplied by 256
3.Bandwidth [10 ^ 7 / (minimum bandwidth link along the path, in the kilobits per second)] * 256
4.Formula with default K values (K1 = 1, K2 = 0, K3 = 1, K4 = 0, K5 = 0):
Metric = [K1 * BW + ((K2 * BW) / (256-load)) + K3 * delay] The
formula does not need to be remembered, you need to understand, which condition needs to be used as the metric value, the corresponding K Set the value to 1, and set it to 0 if it is not needed.
The conditions for EIGRP to establish adjacency :
1. AS number process domain
2.Kvalue
3. Authentication
EIGRP Metric
1. By default, EIGRP uses the following two metrics:
-bandwidth
-delay
2. Not recommended metrics (Cisco The device counts once every 5 minutes by default)
-reliability
-load
Note: Although MTU is exchange in EIGRP packets between neighbor routers, MTU is not factored into the EIGRP metric calculation.
Note: Although MTU performs data exchange in the neighbor table and routing table, but It is still not used as a factor in the calculation of EIGRP metrics.
EIGRP Configurtion
RouterX (config) #router ergip autonomous-system Autonomous system number, process number, you can start multiple EIGRP protocols on a router. If two routes want to establish adjacency, the AS number of EIGRP on both routers must be it's the same
EIGRP also automatically summarizes the network by default and needs to pass
RouterX (config-router) #no auto-summary Turn off automatic summarization
Network announcement: For example, announcement 10.1.1.0/24
1. Main network announcement can be done like RIP
RouterX（config-router）#network 10.0.0.0 255.255.255.0
2. Can also make accurate subnet announcements
RouterX (config-Router) #network 10.1 . 1.0 0.0 . 0.255 (Wild cars subnet mask bits inverted and opposite, on behalf of network 0, bit 0 means 1 denotes a host to match, 1 indicates indifferent)
EIGRP must be accurately notified unless otherwise required.
Cancel the neighbor table update log reminder, but it is not recommended to cancel:
Router（config-router）#no eight log-neighbor-changes，
The routing type of EIGRP displayed in the routing table is D, for example:
D 172.16 . 2.0 [ 90 / 2681856 ] Via 172.16 . 1.2 00 : 01 : 36 , Serial0 / 0
[Management Distance / composite metric]
# managed from two: internal 90 external 170
Some commonly used commands
1. View all agreements
Router #show ip protocol
2. View neighbor table information
Router #show ip eigrp neighbors
H Address Interface Hold(sec) uptime SRTT(ms) RTO Qcnt Seq
H: Established order
Address: Neighbor's IP
Interface: Interface with neighbor
Hold (sec): Default hello timeout time, 3 times the time of hello packet
uptime: Neighbor has been enabled time
SRTT: Smooth Round-Trip Time average round-trip time
RTO ： Retransmit Timeout Retransmit timer
Q cnt (queue count) ： EIGRP In order to prevent the bandwidth of its own protocol bar from being used up, the default protocol traffic only accounts for 50% of the bandwidth. If it exceeds, it will enter the queue. Normally it is 0. If it exceeds 0, it means that the network is congested.
Seq Num: serial number
3. View EIGRP topology table information
Router #show ip eigrp topology
P 10.1 . 2.0 / 24 . 1 Successors, the FD IS 2809856 Via 172.16 . 1.2 ( 2809856 / 2297856 ), Serial0 / 0 .
...
status bits p, ip, 1 successor th, ... (the FD value / AD value), interface
Status: P: passive (passive state, operable state)
A: active (data sending status)
4. View routing table
Router #show ip route
5. View routing information of EIGRP protocol
Router #show ip route eigrp
6. View the relevant interfaces of the EIGRP protocol
Router #show eigrp interface
7. View the statistics of eigrp traffic
Router #show ip eigrp traffic
8.eigrp debugging mode, the data is more complicated, you can understand it, generally not
Router #debug ip eigrp
EIGRP Timer
1. The default EIGRP hello timer on the line with a bandwidth less than T1 is 60s by default, and the hello timer on the T1 line is 5s. The hold-down timer defaults to three times the hello time
. 2. The EIGRP hello time on a common serial interface Yes 5s
3. Modify the convergence time, modify based on the interface, n is the number of seconds
Router（config-if）#ip hello-interval eigrp as号 n
Router（config-if）#ip hold-time eigrp as号 3n
Note: If a modification is made on one interface, all interfaces on the other device must be modified.
EIGRP K Value
By default, EIGRP only uses the two parameters of bandwidth and delay as the metric.
Default K value:
k1 = 1, k2 = 0, k3 = 1, k4 = 0, k5 = 0
Modify the metric, if modified to only Taking bandwidth as a metric, change k3 to 0
Router（config-router）#metric weights 0 1 0 0 0 0
The first 0 is Tos (Type of Service), which does not implement the function and has no practical meaning. It is always set to 0
Note: EIGRP will check the K value when establishing the adjacency relationship. Different routers must have the same K value to establish the adjacency relationship. After
modifying the K value, the router will be down, which will cause network disconnection.
EIGRP Load Balancing
By default, EIGRP uses equal-metric load balancing.
By default, load balancing on up to 4 cost paths
can be configured on up to 16 paths for load balancing: the
maximum number of load balancing entries can be specified through the maximum-paths command
EIGRP unequal-cost unequal-Cost load balancing
Currently (2011) only EIGRP can officially support unequal- cost load balancing
Router (config-router) #variance multiplier (multiple of FD)
Alllows the router to load-balance across routes with a metric smaller than the multiplier value times the minimum metric route to that destination.
The default variance is 1 , which means equal-cost load balancing.
EIGRP MD5 Authentication
function: In order to prevent illegal unauthorized routers or devices from sending illegal routing information to affect the routing table
1. EIGRP only supports MD5 authentication
2. The router checks each received data packet
3. The authentication fails and cannot be established Three conditions for adjacency, adjacency: ①AS number ②K value ③Authentication
4. Each pair of routers must be configured with the same key Key (key value and key-string must be the same)
EIGRP MD5 Authentication Configuration Steps
1. Create key database keychain, create key keys
2. Specify Key ID
3. Configure key-string
4. (Optional) Configure key effective time Specify key effective time
through the following two commands:
Router（config-keychain-key）# accept-lifetime start-time {infinite| end-time |duration seconds} Router（config-keychain-key）# send-lifetime start-time {infinite| end-time |duration seconds}
5. Enable MD5 authentication on the interface
Router（config-if）# ip authentication mode eigrp autonomous-system md5
6. Specify the key database used for authentication on the interface
Router（config-if）# ip authentication key-chain eigrp autonomous-system name-of-chain
Since authentication is specified on the interface, two interfaces of a router can be achieved, one with authentication enabled and one without authentication
EIGRP Route Summarization Automatic
EIGRP can be manually summarized at any point, at any point, it is summarized at the interface level.
1. Purpose: smaller routing tables, smaller updates, smaller routing table information, and routing updates that are less likely to change
Route（config-router）# no auto-summary Route（config-if）# ip summary-address eigrp as-number address mask [admin-distance]
2.automatic summarization：
①On major network boundaries， subnetworks are summarized to single classful（major） network。
②Automatic summarization occurs by default。
EIGRP experiment configuration:
1. View port information:
Router #show interface serial0/0
2. Change the bandwidth:
Router (config) # bandwidth xxxx (unit is kb, the default setting is 1544kb)
3. Accurate protocol channel: use reverse code: such as: 200.1.1.0/24, then bind as follows:
Router（config-router）network 200.1.1.0 0.0.0.255
4. Clear the IP configuration information of the interface:
Router（config）default interface fastethernet0/0
4.1 Expansion, clear configuration of switch range:
default interface range fastethernet0/1-15
5. After the routes are manually summarized, there will also be a summary route on the configured router, which automatically appears and points to the null interface (Null0). All entries that point to the null interface are actually discarded. The reason for this route entry is: to prevent routing black
holes and black hole routes: to suck all irrelevant routes into them, so that they have no back and forth routes, generally route entries that are actively established by admin
Null0 interface: belongs to the logical interface of the router, always It is in the UP state, but does not forward any message. When null0 receives the message, it will discard the message.