Notes on BGP part of CCNP

ospf most 1W Routes
is-is most 2W Routes

www.cidr-report.org

CUII CHINA UNICOM Industrial Internet Backbone, CN

adjacent
Britain [əˈdʒeɪsnt]
beautiful [əˈdʒeɪsnt]
adj. adjacent ; Adjacent to

BGP,border gateway protocol
Border gateway routing protocol , Path vector

single homed Single night , Only connect to one isp, And no redundant links
dual homed Two speed , Only connect to one isp, Two links are used to provide redundancy ,
multihomed multihoming , Connect to multiple isp,
dual multihomed Double sleepover , Connect to multiple isp, At the same time, there are redundant links ,

bgp, Router , Can only , Route it uses , Inform to , Adjacency , In autonomous systems , The equivalent of ,

AS_PATH,
bgp, Autonomous systems , Number , list
One of the path attributes ,
When one bgp route , In a as Inside bgp neighbor , Between , On delivery ,as_path unchanged , Only when it leaves as when ,bgp Border router , Will be as No. is pushed into the route , The location is as_path front ,
Only , When ,bgp route , Pass to ,ebgp Neighbor time ,as_path Will change

bgp Use tcp agreement , The port number is ：179 10110011

bgp peers ,bgp neighbor

peers , Between , In exchange for , Whole ,bgp Routing table

bgp Router , Send only , The incremental 、 Trigger , Without periodic updates ,

bgp, neighbor , There are mainly ,ebgp、ibgp

initial ,bgp The routing table for , It's empty. , Need to load

bgp, Build a neighborhood , stay bgp Routing process , Hand designated ,neighbor, To build neighborhoods ,

route-server.ip.att.net
rviews,rviews

validation It works ;〔 merchant 〕 take effect ; verify
valid （ Legal ） Effective ;（ formal ） Approved ; Logical ; reasonable

[email protected]> show route all

inet.0: 879975 destinations, 14077825 routes (879975 active, 0 holddown, 0 hidden)

• = Active Route, - = Last Active, * = Both

0.0.0.0/0 *[Static/5] 6d 10:23:49
> to 12.0.1.1 via em0.0
1.0.0.0/24 *[BGP/170] 3d 18:15:08, localpref 100, from 12.122.83.238
AS path: 7018 13335 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:23:19, localpref 100, from 12.122.120.7
AS path: 7018 13335 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:23:19, localpref 100, from 12.122.124.12
AS path: 7018 13335 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:23:19, localpref 100, from 12.122.124.67
AS path: 7018 13335 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 2d 17:41:04, localpref 100, from 12.122.124.138
AS path: 7018 13335 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 2d 13:23:05, localpref 100, from 12.122.125.6
AS path: 7018 13335 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:23:19, localpref 100, from 12.122.125.44
AS path: 7018 13335 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 4d 16:49:21, localpref 100, from 12.122.125.106
AS path: 7018 13335 I, validation-state: valid
—(more)—
[BGP/170] 6d 10:23:00, localpref 100, from 12.122.125.132
AS path: 7018 6939 10089 9587 24378 24378 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:04:46, localpref 100, from 12.122.125.165
AS path: 7018 6939 10089 9587 24378 24378 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:23:00, localpref 100, from 12.122.125.224
AS path: 7018 6939 10089 9587 24378 24378 I, validation-state: valid
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:23:00, localpref 100, from 12.122.126.9
—(more)—

                [BGP/170] 6d 10:24:03, localpref 100, from 12.122.126.9
                  AS path: 7018 2914 9605 I, validation-state: unknown
                > to 12.0.1.1 via em0.0
                [BGP/170] 6d 10:23:56, localpref 100, from 12.122.126.64
                  AS path: 7018 2914 9605 I, validation-state: unknown
                > to 12.0.1.1 via em0.0
                [BGP/170] 6d 10:24:07, localpref 100, from 12.122.126.232
                  AS path: 7018 2497 9605 I, validation-state: unknown
                > to 12.0.1.1 via em0.0

—(Head of output truncated; more)—

[BGP/170] 3d 20:57:44, localpref 100, from 12.122.159.217
AS path: 7018 2914 9318 55625 I, validation-state: unknown
> to 12.0.1.1 via em0.0
1.248.0.0/13 *[BGP/170] 2d 20:20:28, localpref 100, from 12.122.83.238
AS path: 7018 6453 9318 9318 I, validation-state: unknown
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:24:22, localpref 100, from 12.122.120.7
AS path: 7018 6453 9318 9318 I, validation-state: unknown
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:24:17, localpref 100, from 12.122.124.12
AS path: 7018 6453 9318 9318 I, validation-state: unknown
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:24:12, localpref 100, from 12.122.124.67
AS path: 7018 6453 9318 9318 I, validation-state: unknown
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:23:56, localpref 100, from 12.122.124.138
AS path: 7018 6453 9318 9318 I, validation-state: unknown
> to 12.0.1.1 via em0.0
[BGP/170] 6d 10:23:42, localpref 100, from 12.122.125.6
AS path: 7018 6453 9318 9318 I, validation-state: unknown
—(Head of output truncated; more)—

[email protected]> show route summary
Autonomous system number: 65000
Router ID: 12.0.1.28

inet.0: 879958 destinations, 14077560 routes (879958 active, 0 holddown, 0 hidden)
Direct: 1 routes, 1 active
Local: 1 routes, 1 active
BGP: 14077441 routes, 879839 active
Static: 117 routes, 117 active

inet6.0: 143535 destinations, 2296243 routes (143535 active, 0 holddown, 0 hidden)
Direct: 1 routes, 1 active
Local: 2 routes, 2 active
BGP: 2296237 routes, 143529 active
Static: 2 routes, 2 active
INET6: 1 routes, 1 activesh

BGP,
l2 header,ip header,tcp header,bgp packets,FCS
6 port：179 BGPheader,BGPpacket
open,update,notification,keepAlive,route-refresh
send out keepAlive, Waiting for each other's keepAlive,

expiry expire

function BGP The router , This is called BGP speaker,
BGP neighbor , Also called ,BGP peer

peer peers ; contemporary ; identity （ Or status ） The same people

peers=neighbor
EBGP,external BGP, external BGP Relationship ,
IBGP,internal BGP, Inside BGP Relationship ,

internal Inside ; Inside ;
international international

external External ; Outside ;

EBGP, It is generally used , Native Driver , Default ttl=1

A router , Can only , Activate , One BGP, process ,

Support one-time updating of a large number of routing entries .

Send periodically Keepalive Information to verify TCP Whether the connection is normal

,OSPF An update 100 Routes

As soon as the neighborhood is established ,BGP Speaker Just take what you have BGP The optimal route is announced to the new neighbor

keepalive It's a 19 byte , Sent periodically BGP Message header , No data fields

Update（code 2）： The message contains three components ： Network Layer Reachability messages （NLRI）、 Path properties and revoked routes .

Update You can delete （ Declare inaccessible ） And increase （ Declare reachable ） route . The content is the length of the prefix .

BGP Of 3 A database
Neighborhood watch 、BGP Forwarding table （ Also called forwarding library ）、IP Routing table

BGP It's a kind of AS-by-AS The routing protocol of , Its next jump refers to the next AS. Not the next Router

from EBGP Learned routing , Its next hop attribute remains unchanged

Its perspective is AS, Not a single one router

http://www.jdccie.com/?p=3441

edge AS, transmission AS（ transit AS）,

PS C:\Users\username> net user username 20222022
The password does not meet the requirements of password policy . Check the minimum password length 、 Password complexity and password history requirements .

Please type the NET HELPMSG 2245 To get more help .

PS C:\Users\username> net user username Qwer2022

BGP Sync ,
IGP synchronization

igp and bgp Synchronization of , Is refers to , from ibgp Received route , Also through igp Can be up to , Otherwise, I won't give ebgp neighbor ,
This is because , If from ibgp Received route , But not through igp arrive , This leads to a routing black hole , Also let ebgp neighbor
Cannot reach by oneself ibgp The destination network segment in the incoming route , therefore , Will not use igp Unreachable route to ebgp neighbor
This is it. igp and bgp Synchronization of the relevant knowledge points ,

BGP surface , From the neighbors , All the routes , The destination network segment will have , Multiple paths ,
Routing table , Lists the... That arrive at the destination network segment , The only way , That is to say best route ,BGP No load balancing ,
EBGP Manage distance ,AD administrator distance,20

BGP Basic configuration
1、 establish BGP process ,
router bgp autonomous-system
Only one router can be configured BGP example , in other words , A router can only belong to one as, It is based on the whole router , Not by interface ,

router bgp 100
router bgp 200
bgp is already running;as is 100

2、 Appoint BGP neighbor , Activate BGP conversation
neighbor{ip-address|peer-group-name} remote-as autonomous-system
Neighbor's ip Address
Neighbors must be accessible ,
Can only be specified manually ,（ Because it's point-to-point unicast , Do not send multicast , Therefore, automatic discovery cannot be realized ）
as It's a direct decision , The neighbors are EBGP conversation , still IBGP conversation ,

neighbor 12.1.1.1 remote-as 100

Under process ,bgp router-id x.x.x.x, Manual settings bgp router-id, It will appear in open In the message ,

IBGP, When building neighborhoods , Interface addresses are not recommended , because IBGP May not be directly connected , such , There may be
The same network segment exists on the path , Then, two interfaces have the same address , Unable to determine the update source update-source, therefore , Don't suggest
Use the interface address to establish a neighbor relationship , But use ,router-id To build neighborhoods ,

IBPG, Build a neighborhood , Recommended ,loopback Interface as （ Instead of using the address of the interface ）, Source , To build BGP neighbor ,
neighbor 3.3.3.3 update-source loopback0

Here's why ：
hostname R3
router bgp 123
neighbor 10.1.13.1 remote 123
neighbor 10.1.23.2 remote 123

stay IBGP In the environment of , Routing reachability is determined by IGP be responsible for ,IGP There are often routing redundancy settings ,
therefore , It is better to use the interface than to use loopback mouth
In this case , When normally established BGP After the link between neighbors is down ,BGP Will automatically try to use IGP Provide redundant paths to establish neighbors

By grabbing bags ,keepalive In bag , You can see update-source Update source , It can also be done through neighbor 3.3.3.3 update-source loopback0
To change the update source , That's the source address , The source address is local ,
When not used ,neighbor 3.3.3.3 update-source loopback0, To specify when to update the source ,
The address of the update source is , Go to 3.3.3.3 Address of the local outgoing interface of ,
When using the ,neighbor 3.3.3.3 update-source loopback0, after
The address of the update source is ： The local loopback0 Address ,

show tcp brief tcp Connect , Abstract
show ip bgp bgp surface
show ip bgp rib-failure Not loaded into rib Medium bgp route , And why
show ip bgp neighbors bgp Neighborhood watch , Details
show ip bgp summary summary of bgp neighbor status
clear ip bgp * soft Ruanqing ,bgp Will not be disconnected

Supply chain management (Supply Chain Management , abbreviation SCM):

Appoint ,bgp The network that will be notified
network network-number [mask networkmask] [route-map map-tag]

Of the party “ Six disciplines ” It's political discipline 、 Organizational discipline 、（ Integrity and discipline ）、 Mass discipline 、（ Work discipline ）、 Life discipline .

There is no sign > Of bgp route , It is not an optimal route ,
1、 Do not put it into the routing table
2、 It will not be passed on to bgp neighbor

bgp Conditions for route optimization ：
1、 Sync , from igp Learning to / Write static / Turn off the synchronization function
2、 Next jump , from igp Learning to / Write static /net-hop-self

bgp, yes ,as-by-as Of , Routing protocol , instead of router-by-router, The routing protocol of ,
stay bgp in ,next-hop, It refers to the next arrival as Interface address in , Not the next router ,
ebgp in ,next-hop, It's the next one as Of , Send updates to the neighbor router ip Address ,
from ebgp From the next-hop attribute , stay ibgp In the middle of the transmission , remain unchanged ,

neighbor it next-hop-self
ibgp Manage distance 200
ebgp Manage distance 20

show ip bgp neighbor {address} received-routes
show ip bgp neighbors {address} routes
show ip bgp neighbors {address} advertised-routes

show ip protocol
routing protocol is “ospf 110”
maximum path：4

show ip protocol
routing protocol is “bgp 200”
maximum path：1

NLRI,network layer reachability information

path attribute
1、origin
2、as_path
3、next_hop
4、multi_exit_disc

Attribute classification
1、 Recognized attributes well-known
2、 Optional attribute optional

Recognized attributes ：
1、 Recognized mandatory properties well-known mandatory
2、 The accepted attribute of free decision well-known discretionary

Optional attribute ：
Optional and transitive optional transitive
Optional but not transitive optinal non-transitive

Recognized attributes ： all bgp Router , Can identify and support
mandatory Must appear in update In the message , namely update The message must have it
discretionary update The message does not necessarily have it

mandatory
Britain [ˈmændətəri]
beautiful [ˈmændətɔːri]
adj. Compulsory ; Compulsive ; Statutory ; Obligatory
discretionary
Britain [dɪˈskreʃənəri]
beautiful [dɪˈskreʃəneri]
adj. Freely determined ; Acting at discretion

Optional attribute ：bgp Router , Not all are recognized and supported
transitive Even if it doesn't support , Also pass it on to others bgp neighbor
non-transitive If not , They will not pass on , （ Maximum degree of freedom ）

med,multi_exit_disc

https://blog.csdn.net/weixin_43394724/article/details/120781463

network layer reachability informaiton
Is specific bgp Route entry

discriminator
Britain [dɪsˈkrɪmɪneɪtə]
beautiful [dɪˈskrɪməˌneɪtər]
n. Discriminator , Discriminator

“MED(Multi-Exit Discriminators) Property is an optional non transitional property .

One 、weight- Optional non delivery
Cisco private , Locally available , The local policy , Not to any bgp neighbor
Local origin , The default weight is ：32768
From the other bgp Neighbors learn from ：0
weight, The bigger the better // Positive solution

Two 、local preference- It is recognized that self-determination
A preference for ; hobby ; love ; Favorite things
Local priority ,
Only in as Intransition , namely ibgp Inter transmission , Not to others as, Do not give ebgp neighbor
From other ibgp neighbor , Only to ibgp neighbor
inform , Ben as, Leave as Of , Preferred path , come from as Internal bgp The neighbor role is ebgp // Positive solution
Default ：100
from as External ebgp Routes learned by neighbors ,local preference It's empty
local preference, The higher the better
bgp default local-preference 500 // modify local preference value

3、 ... and 、as-path- It is acknowledged that it must comply with / Recognized compulsion
as_set disorder
as_sequence Orderly
as_confed_sequence Orderly Federation //confederacy union ; alliance ; The federal
as_confed_set Disordered Federation

type=as sequence
value=300 100

Four 、origin- It is acknowledged that it must comply with / Recognized compulsion
-i come from bgp network, Come of igp
-？ Routes from other channels, such as republishing ,
i>incomplete

in incoming-routes
out outbound-routes

Use route-map modify origin
1、ip prefix-list 1 seq 5 permit 4.4.4.0/24
2、route-map rpname permit 10
match ip address prefix-list 1
set origin egp 200[remote as]
route-map rpname permit 20
3、router bgp 100
neighbor 24.1.1.4 route-map rpname in [incoming-routes,outbound-routes]
4、clear ip bgp * soft

5、 ... and 、MED Multi-Exit Discriminators- Optional non delivery
Only in two as Between , swapping ,
measurements , The smaller the better
Conditional comparison , Only if the opposite as Inside , There are two ebgp When neighbors , To compare med
To external neighbors , Pointed out that , Enter Ben as Of , Preferred path ,
Affect the opposite end as choice , Enter Ben as The path of ,
exit exit ; way out
（local preference Is to leave Ben as Preferred path for ,MED Tell the other person as Enter Ben as Preferred path for 【 other party as How many exits 】）
The default value is ：0
Locally originated bgp route ,network/redistribute, to ebgp When neighbors , Need to carry MED,
to ibgp When neighbors , Be sure to bring MED
network, Local , Native Driver , Enter into bgp in ,med=0
network, Local , Static routing , Get into bgp in ,med=0

6、 ... and 、next_hop- It is acknowledged that it must comply with
Issues related to accessibility , Such as extraterritorial bgp Route entry , The required roles are ebgp Of bgp Router usage next-hop self Parameters are notified to it ibgp neighbor ,

7、 ... and 、community- Optional transitive
RFC Format ,AA:NN
2 Bytes of AA, identification as Number
2 Bytes of NN, identification DIY The number of
stay RFC Format and Cisco Commands for switching between formats ,ip bgp-community new-format
neighbor X.X.X.X send-community

8、 ... and 、automic_aggregate
aggregator

aggregate-address 172.16.0.0 255.255.0.0 summary-only
summary-only Suppress details
aggreated by 300 3.3.3.3

BGP Automatic summarization
auto-summary
Reissue , Will be summarized
network, Still update with details ,

BGP The opening configuration trilogy ：
1、no synchronization close igp Sync
2、no auto-summary Turn off auto rollup
3、bgp router-id 4.4.4.4 Manual configuration router-id

Manual summary
aggregate-address 172.16.0.0 255.255.0.0 summary-only as-set

A prefix list can correspond to multiple prefixes
ip prefix-list plname permit 172.16.10.0/24
ip prefix-list plname permit 172.16.11.0/24

route-map rmname permit 10
match ip address prefix-list plname

regular expression

General characters
Control characters （ Of special significance ,metaCharacters）
1、 Atomic characters , placeholder 、 control 【.^$_|\】
2、 Multiplicative character , Repetition mode
3、 Range character , Limit the scope

. Represents a single , arbitrarily , General characters , It can also represent a space .any It means , Whether it's a regular character or a control character
^ Start with who , example ^Eth
$ By whom , example sec$
_ Represents a single , arbitrarily , Separator ,^$ ,
| ① Logic or ,② Pipe,
\ Used to put the following , Control characters , Turn into , Ordinary character , Play a , The function of escape

^a.$ With a At the beginning , Any single character , At the end of the , character string .（. Not an extension , It means a single arbitrary character ）
^100_ With 100 At the beginning ,100 There's one in the back , Separator （ arbitrarily ）. Satisfy 100 start , There is a separator , These two conditions .
^100$ With 100 start , With 100 end , So it means ,100.
100$|400$ With 100 end , Or with 400 At the end of the , character string
^(65000)$ It just means ,（65000）

• The character before it , appear 0 Time , Or many times

• The character before it , appear 1 Time , Or many times
  ？ The character before it , appear 0 Time , or 1 Time

abc*d c Is there one , But there is no c, Or more than one c
abc+d c There has to be one , Or more than one c
abc?d c Is there one , There can be no c, But there can't be more c
a(bc)?d bc As a whole, there can be 1 Time ,bc As a whole can not appear

[] Range , Any one of the ranges appears , It's all a match

[^act]$ Do not use a At the end of the , Or don't use c At the end of the , Or don't use t At the end of the ,

community,
1、rfc Format ,ip bgp-community new-format, Is a colon separated decimal form ,
2、16 Base format , What you see in the bag
3、Cisco ios Format , stay Cisco What you see using the command on the device , Is a two in one decimal format

community, It can be used route-map To set up
route-map test permit 10
set community local-as

set community ？
1-4294967295 community number
aa：nn community number in aa：nn format
additive add to the existing community
Internet Internet（well-known community）
local-as do not send outside local as（well-know community）
no-advertise do not advertise to any peer（well-know community）
no-export do not export to next as （well-know community）
none no community attribute

additive Insert , additional , Not cover ,
Internet be-all bgp route , All default to Internet attribute

The default is not to pass community, You need to turn on the delivery ,
neighbor X.X.X.X send-community

local-AS: Routes with this attribute will not be sent out AS( There will be no transmission between federations , There will also be transfers within the Federation , If there is no Federation, it is only in this AS Pass on ) remarks : If I didn't become a federal no-export and local-as The effect is the same

router reflector
Routing reflectors
Only aim at ibgp scene

Route origin ,originator_id
Reflection cluster id list ,cluster_list
These two properties , It will only appear in ,router reflector Environment ,

R3--------R4--------R5
client RR non-client

origintor_id：3.3.3.3
cluster_list：4.4.4.4
If cluster_list, Not specified manually , By default , It's a routing reflector RR Of router-id

Reflector , received cluster_list There's one of them router_id when , For the purpose of environmental protection , This route will not be reflected ,

cluster id, Generally, the RR Of id

Use ,prefix-list To filter out 12.12.12.0/24, Other releases
ip prefix-list 12 deny 12.12.12.0/24
ip prefix-list 12 permit 0.0.0.0/0 le 32
router bgp 12
neighbor 10.1.23.3 prefix-list 12 out

Use ,distribute-list + ACL. To filter out 12.12.12.0/24, Other releases
access-list 1 deny 12.12.12.0
access-list 1 permit any
router bgp 12
neighbor 10.1.23.3 distribute-list 1 out

In the use of distribute-list When , If in neighbor In the sentence , It can only be associated with ACL
<1-199> ip access list number
<1300-2699> ip access list number (expanded range)
word ip access-list name

Use ,distribute-list +prefix-list. To filter out 12.12.12.0/24, Other releases
ip prefix-lsit 12 deny 12.12.12.0/24
ip prefix-lsit 12 permit 0.0.0.0/0 le 32
router bgp 12
distribute-list prefix-list 12 out
out It is not allowed to follow the interface , Otherwise, the policy is invalid ,
This is because , The configured policy , Will be applied to the router , All interfaces , This policy will be applied whenever you go out of the router ,
So when the interface is specified , The policy becomes invalid , Applying to all interfaces means applying to all bgp Neighbors
This one doesn't neighbor The sentence of , Is to be applied to this router , On all interfaces , So we can cooperate with prefix-list Use it together ,

<1-199> ip access list number
<1300-2699> ip access list number (expanded range)
word ip access-list name
gateway filtering incoming updates based on gateway
prefix filtering prefixes in routing updates

in/out, The routing protocol name is not followed , It means , For all routing protocols ,

network 11.11.11.0 mask 255.255.255.0 route-map rpname

R1,network Set the path attribute when introducing a route
ip prefix-list 11 permit 11.11.11.0/24
ip prefix-list 12 permit 12.12.12.0/24
route-map rp1 permit 10
set community 100：11
route-map rp2 permit 20
set community 100：12
router bgp 100
network 11.11.11.0 mask 255.255.255.0 route-map rp1
network 12.12.12.0 mask 255.255.255.0 route-map rp2
neighbor 10.1.12.2 send-community

R1, Yes R2 When routing , Set up MED Property value
ip prefix-list 11 permit 11.11.11.0/24
ip prefix-list 12 permit 12.12.12.0/24
route-map rpname permit 10
match ip address 11
set metric 1000
route-map rpname permit 20
match ip address 12
set metric 2000
route-map rpname permit 30 // Release the remaining routes , Otherwise, only the above two can be sent to the neighbors
router bgp 100
neighbor 10.1.12.2 route-map rpname out
neighbor 10.1.12.2 send-community

route-map, There is one at the bottom by default deny any The sentence of , Implicitly reject all ,
route-map To write an empty statement , Otherwise, all... Will be rejected by default

O intra-area In the area
OIA inter-area Interregional

confederation union ; Consortium

confederation（ The federal as） Contains several... Using private numbers member（ member as）, For others as Come on , They are unknown members as The presence of the ,

1、router bgp member
2、 Where the declaration is located confederation
3、member Between Mutual reference peer, If there is confederation peer Also point out

R1--------R3--------R4--------R5--------R2
as100 as345 as345 as345 as200

R3：
router bgp 64512
bgp confederation identifier 345
neighbor 4.4.4.4 remote-as 64512
neighbor 4.4.4.4 update-source loopback 0
neighbor 10.1.13.1 remote-as 100

R4：
router bgp 64521
bgp confederation identifier 345
bgp condederation peers 64513
neighbor 3.3.3.3 remote-as 64512
neighbor 3.3.3.3 update-source loopback 0
neighbor 5.5.5.5 remote-as 64513
neighbor 5.5.5.5 ebgp-multihop 3 // Use loopback to establish ebgp Relationship , Note the number of hops , Under normal circumstances, the loopback port to the router is a hop
neighbor 5.5.5.5 update-source loopback 0
（ bgp condederation peers 64513 // Indicate the Federation ebgp Neighbor's as）

loopback--------R4--------R5--------loopback
+1 +1
hop hop

ebgp Our neighbors set up ：
1、 Use it directly , Native Driver ,
2、 Use loopback interface , But pay attention to the problem of multi hop ,multihop

R5：
router bgp 64513
bgp confederation identifier 345
bgp condederation peers 64512
neighbor 4.4.4.4 remote-as 64512
neighbor 4.4.4.4 ebgp-multihop 4
neighbor 4.4.4.4 update-source loopback 0
neighbor 10.1.25.2 remote-as 200
——————————————————————————
R1：
show ip int br
show run | s r o
router ospf 100
router-id 1.1.1.1
network 1.1.1.0 0.0.0.255 area 0
network 12.1.1.0 0.0.0.255 area 0
network 13.1.1.0 0.0.0.255 area 0
show run | s r b
router bgp 64512
bgp router-id 1.1.1.1
no synchronization
no auto-summary
bgp confederation identifier 100
neighbor 2.2.2.2 remote-as 64512
neighbor 2.2.2.2 update-source loopback 0
neighbor 3.3.3.3 remote-as 64513
neighbor 3.3.3.3 update-source loopback 0
neighbor 3.3.3.3 ebgp-multihop 2 //ebgp-multihop ,multihop Is the wrong abbreviation
bgp confederation peers 64513
do show ip bgp summary
show ip bgp
show tcp brief
clear ip bgp * soft

R3：
router bgp 64513
bgp router-id 3.3.3.3
no synchronization
no auto-summary
bgp confederation identifier 100
neighbor 1.1.1.1 remote-as 64512
neighbor 1.1.1.1 update-source loopback 0
neighbor 1.1.1.1 ebgp-multihop 2
bgp confederation peers 64512
neighbor 35.1.1.5 remote-as 300
neighbor 35.1.1.5 update-source loopback 0
neighbor 1.1.1.1 next-hop-self

R5：
router bgp 300
bgp router-id 5.5.5.5
no synchronization
no auto-summary
neighbor 35.1.1.3 remote 100
network 5.5.5.0 mask 255.255.255.0
do show ip bgp

R2：
show run | se router
router ospf 110
router-id 2.2.2.2
network 2.2.2.0 0.0.0.255 area 0
network 12.1.1.0 0.0.0.255 area 0
（router bgp 100
no synchronization
no auto-summary
bgp router-id 2.2.2.2
neighbor 3.3.3.3 remote-as 100
neighbor 3.3.3.3 update-source loopback 0
neighbor 3.3.3.3 next-hop-self
neighbor 24.1.1.4 remote-as 200 ）
router bgp 64512
bgp router-id 2.2.2.2
no synchronization
no auto-summary
bgp confederation identifier 100
neighbor 1.1.1.1 remote-as 64512 // here , I need to look at it first ,1.1.1.1 Whether it has passed ospf Can be up to
neighbor 1.1.1.1 update-source loopback 0
neighbor 24.1.1.4 remote 200
do show run | s r b
neighbor 1.1.1.1 next-hop-self
neighbor 1.1.1.1 update-source loopback 0 The function of is to neighbor 1.1.1.1 next-hop-self To prepare for , Do matting

R4：
router bgp 200
bgp router-id 4.4.4.4
neighbor 24.1.1.2 remote 100
do show ip int br
net 4.4.4.0 mask 255.255.255.0
do show ip bgp

stay bgp in ,13 In the skip routing rule ,1_{3- The higher the value, the better ,4}13- Small value is better

In the current network , Common routing rules ,lp、med

1、weight, Maximum , optimization
2、local_preference, optimization , Maximum
3、 Local origin next-hop 0.0.0.0, optimization
4、as-path, The shortest , optimization
5、origin,igp>egp>incomplete
6、med, Minimum , Only if all alternatives come from the same as,
7、ebgp The neighbor who sent it > from ibgp The neighbors learned it , In the Federation , Federation is preferred ebgp
8、 To bgp next_hop, The nearest route
9、bgp Load balancing
10、 The one who lives the longest ebgp neighbor ,
11、router-id, Minimum
12、cluster-list, The shortest
13、 Neighbor address , Minimum .neighbor The neighbor address configured in the command

weight,32768
neighbor A weight 100
neighbor B weight 200
Modify the weight of all routes received from a specific neighbor

bgp default local-preference , Default 100

show ip bgp
show ip bgp 100.1.1.0/24

Set locally declared 8.0.0.0 The network segment weight The value is 0
route-map lp20 permit 10
set weight 0
router bgp 345
network 8.0.0.0 mask 255.0.0.0 route-map lp20

traffic share count is 1
The equivalent load balancing traffic is ,1 Than 1, Nonequivalent load balancing is not 1 Than 1 了 .