This document describes the configuration of a network with Juniper and Cisco routers using OSPF, iBGP with route reflectors, and eBGP. It involves: 1. Configuring OSPF in area 0 between the Juniper and Cisco routers. 2. Configuring the Juniper routers J03 and J04 as iBGP route reflectors to allow the exchange of iBGP routes between the left and right networks. 3. Adding the Cisco routers R1 and R2 as iBGP peers to J03 to learn routes from the left network, and similarly configuring R3 and R4 as iBGP peers to J04 to learn routes from the right network.

1. 1. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco – Basics Setup & OSPF
Going back to basics with Network Design, Route-Reflectors (iBGP), OSPF & Finally eBGP
J3 – Loop 10.0.0.3
J4 – Loop 10.0.0.4
R1 – Loop 10.0.0.10
R2 – Loop 10.0.0.11
R3 – Loop 10.0.0.12
R4 – Loop 10.0.0.13
J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
1. Get OSPF Talking in Area 0
2. Get J03 – J04 Talking iBGP-RR
3. Add R1/R2 → iBGP – J03 (left)
4. Add R3/R4 → iBGP – J04 (right)
5. Add eBGP PEERS
6. Design Route-Maps OUT
7. Design some Resiliency
AS99
My IP ADDRESS Range
10.0.0.0/24 – Reserved for Loopbacks
20.0.0.0/30 – Private links betw J03-J04
30.0.0.0/24 – Reserved for ibgp (left)
40.0.0.0/24 – Reserved for ibgp (right)
left right
Have a closer look at our Network AS99. It really looks like two networks
separated only by the 20.0.0.0/30 portion (interface em1). In fact, we can
imagine that the “left” was the first network and later after expansion, a “right”
network as added together with a the new IP Range 40.0.0.0/24. - Later on we
will consider some “problems” with this design and how perhaps to overcome it.
Junos J03
kjteoh@Junos-3> show configuration interfaces lo0
unit 0 {
description Loop0;
family inet {
address 10.0.0.3/32;
}
kjteoh@Junos-3> show configuration interfaces em1
description Junos3-Junos4;
unit 0 {
family inet {
address 20.0.0.1/30;
}
kjteoh@Junos-3> show configuration routing-options router-id
router-id 10.0.0.3;
kjteoh@Junos-3> show configuration protocols ospf
area 0.0.0.0 {
interface all;
Cisco R1/R2
interface Loopback0
description loop
ip address 10.0.0.10 255.255.255.255
!
interface FastEthernet0/1
description R1-Junos3
ip address 30.0.0.2 255.255.255.0
!
router ospf 99
router-id 10.0.0.10
log-adjacency-changes
redistribute connected subnets
passive-interface default
no passive-interface FastEthernet0/1
network 0.0.0.0 255.255.255.255 area 0
Make sure everything works.
kjteoh@Junos-3> show ospf neighbor
Address Interface State ID
30.0.0.2 em0.0 Full 10.0.0.10
30.0.0.3 em0.0 Full 10.0.0.11
20.0.0.2 em1.0 Full 10.0.0.4
R2#show ip ospf neighbor
Neighbor ID Pri State Dead Time Addres
10.0.0.3 128 FULL/DR 00:00:39 30.0.0.1
10.0.0.10 1 FULL/BDR 00:00:38 30.0.0.2
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
Why is this important?
Answer: the Cisco router are on the edge
and if it is connected to an external peer the
Cisco router will try and do OSPF with it and
send hello packets. Not good!

2. 2. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco – Config RR
Making J03 and J04 participate as Route Reflector & iBGP neighbors
J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
Really important to plan stuff out here especially with Junos. You have the options to
create bgp “groups” which is an advantage but can go haywire/
messy/unmanageableif you don't properly plan it in the configs.
J03 iBGP Configs & RR
kjteoh@Junos-3> show configuration protocols bgp group ibgp-RR-ONLY
type internal;
local-address 10.0.0.3;
advertise-peer-as;
family inet {
unicast;
}
export ibgp_export;
cluster 10.0.0.3;
local-as 99;
neighbor 10.0.0.4 {
description Junos4-RR;
}
Create a “group” and stick with it
IBGP ONLY
Define your Cluster
You're DONE!
Repeat the same on J04
cluster 10.0.0.4
Create Route-Map “out”
kjteoh@Junos-3> ...cy-options policy-statement ibgp_export
term 1 {
from protocol direct;
then accept;
}
term 3 {
from protocol static;
then accept;
}
term 2 {
from protocol ospf;
then accept;
Mostly self-explanatory but
this rule is interesting and it
had to be created to make a
40.0.0.0/24 network (right)
available to the 30.0.0.0/24
network on the left!
(In iBGP)
How do Route-Reflectors work?
First, this only applies to iBGP. R1 & R2 only need to do iBGP with J03. J03 in
turn learn iBGP routes from J04 and tell J03 about them. J04 will have his own set
of iBGP neighbors … in our case, R3 & R4. They will learn routes from the “left”
network via the exchange from J03 ↔ J04.
Of course it is also possible for the Cisco clients R1 & R2 to do ibgp with J04 and it
is a good idea too (dotted blue). J03 might fail. If this is the case, it is best to
ensure that the Left network can physically find its way to the Right network. This
can be achieved by trunking the switches above.
But we will also understand that OSPF adjacency will grow for ALL the routers as
they will become direct neighbors. Something to take note of.
BUT, there is a DOWNSIDE to having Route-Reflectors over FULL-MESH. Can
anybody tell me what it is?
IBGP-LEFT
Trunk
IBGP- RIGHT .. Maybe Later
kjteoh@Junos-3> show route receive-protocol bgp 10.0.0.4
inet.0:
Prefix Nexthop MED Lclpref AS path
10.0.0.4/32 10.0.0.4 100 I
10.0.0.12/32 40.0.0.2 2 100 I
10.0.0.13/32 40.0.0.3 2 100 I
20.0.0.0/30 10.0.0.4 100 I
40.0.0.0/24 10.0.0.4 100 I
Loops & interface IP
of R3 & R4 from
“right” network
“Right” network origination
the 40.0.0.0/24 block
kjteoh@Junos-3> show route table inet.0 40.0.0.0/24
40.0.0.0/24 *[OSPF/10] 01:14:33, metric 2
> to 20.0.0.2 via em1.0
[BGP/170] 01:13:43, localpref 100, from 10.0.0.4
AS path: I
> to 20.0.0.2 via em1.0

3. 3. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco – Config RR – Adding iBGP Peers R1 & R2
Adding iBGP Peers to J03... R1 & R2 – Prepare J03 to accept clients
J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
Configs J03
group ibgp {
type internal;
local-address 10.0.0.3;
advertise-peer-as;
family inet {
unicast;
}
export ibgp_export;
cluster 10.0.0.3;
local-as 99;
neighbor 10.0.0.10 {
description Cisco-R1;
}
neighbor 10.0.0.11 {
description Cisco-R2;
}
}
Specific group
created for Left
network
RR cluster
ID for J03 &
local AS99
Neighbor
IP
Configs
Route-Map
OUT
Configs R1/R2
router bgp 99
neighbor 10.0.0.3 remote-as 99
neighbor 10.0.0.3 update-source Loopback0
!
address-family ipv4
neighbor 10.0.0.3 activate
neighbor 10.0.0.3 soft-reconfiguration inbound
!
Neighbor IP
Loop J03
IBGP
Is this Important? Does
R1 & R2 have multiple
exit points to J03?
What about Next hop-
self?
What is soft-recon ..
Is this mandatory?
Properly configured …
J03
kjteoh@Junos-3> show bgp summary
Groups: 3 Peers: 5 Down peers: 0
Peer AS InPkt OutPkt OutQ
10.0.0.4 99 87 88 0 0
10.0.0.10 99 82 97 0 0
10.0.0.11 99 81 96 0 0
Cisco R2
R2#show bgp sum
BGP router identifier 10.0.0.11, local AS number 99
Neighbor V AS MsgRcvd MsgSent Tbl
10.0.0.3 4 99 88 76 24
My IBGP AS
This is J04
configured
in slide 2
The Cisco
R1 & R2
neighbor
The J03 neighbor configured. I
will learn router R1 from this
iBGP neighbor (not ospf).
Does this mean that I (R2) will
INSTALL it?
R2#show ip bgp neighbors 10.0.0.3 routes
BGP table version is 24, local router ID is 10.0.0.11
Origin codes: i - IGP, e - EGP, ? - incomplete
Network Next Hop Metric LocPrf Weight Path
*>i10.0.0.0/24 10.0.0.3 100 0 i
r>i10.0.0.3/32 10.0.0.3 100 0 i
r>i10.0.0.4/32 20.0.0.2 1 100 0 i
r>i10.0.0.10/32 30.0.0.2 2 100 0 i
In terms of ROUTING,
which prefix will be
installed in Cisco R2?
10.0.0.0/24 or
10.0.0.10/32?
Compare: show ip route
10.0.0.10
Why is this
MANDATORY?
Cisco syntax
below

4. 4. Back to Basics – OSPF / iBGP (Route Reflectors) - What R1/R2 is really learning from “Right” side
Refer slide 3 and perform the same between J04 ↔ Cisco R3 & R4. We will have the
following setup … One of the main Questions you should be asking yourself is … who and
which router is ORIGINATING routes for ..10.0.0.0/24, 30.0.0.0/24 & 40.0.0.0/24?
How are routes originated? Is this an automatic process? Is this OSPF or BGP?
Is there such a thing as originating routes in OSPF?
J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
Left Network J03
Junos03 ibgp neighbors
R1 @ 10.0.0.10
R2 @ 10.0.0.11
Router J03
Right Network J04
Junos04 ibgp neighbors
R3 @ 10.0.0.12
R4 @ 10.0.0.13
Router J04
IBGP
20.0.0.0/30
Router R1
Router R2
IBGP
10.0.0.0/24
30.0.0.0/24
IBGP
10.0.0.0/24
40.0.0.0/24
Router R3
Router R4
Cluster RR
10.0.0.3 & .4
R2#show ip route
20.0.0.0/30 is subnetted, 1 subnets
O 20.0.0.0 [110/11] via 30.0.0.1, 01:54:57, FastEthernet0/1
40.0.0.0/24 is subnetted, 1 subnets
O 40.0.0.0 [110/12] via 30.0.0.1, 01:50:53, FastEthernet0/1
10.0.0.0/8 is variably subnetted, 7 subnets, 2 masks
O 10.0.0.10/32 [110/11] via 30.0.0.2, 01:56:32, FastEthernet0/1
C 10.0.0.11/32 is directly connected, Loopback0
O 10.0.0.12/32 [110/13] via 30.0.0.1, 01:50:55, FastEthernet0/1
O 10.0.0.13/32 [110/13] via 30.0.0.1, 01:50:55, FastEthernet0/1
O 10.0.0.3/32 [110/10] via 30.0.0.1, 01:55:00, FastEthernet0/1
B 10.0.0.0/24 [200/0] via 10.0.0.3, 01:54:19
O 10.0.0.4/32 [110/11] via 30.0.0.1, 01:51:17, FastEthernet0/1
30.0.0.0/24 is subnetted, 1 subnets
C 30.0.0.0 is directly connected, FastEthernet0/
Relevant Codes:
B – BGP
C – connected
O – OSPF
This is a small network and we should really take
the time to go through every route and understand
how it is learned and where it is coming from.
We can “learn” routes from many Routers &
SOURCES (protocols), but we install only ONE
route and use it for routing.
It is possible to learn one route from one router and
from different source (protocols).
R2#show ip route 10.0.0.2
Routing entry for 10.0.0.0/24
Known via "bgp 99", distance 200, metric 0, type internal
Last update from 10.0.0.3 00:00:08 ago
Routing Descriptor Blocks:
* 10.0.0.3, from 10.0.0.3, 00:00:08 ago
Route metric is 0, traffic share count is 1
AS Hops 0
This becomes really
important when you
start doing EBGP with
other networks / ASNs

5. 5. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco – Adding EBGP Peers / Neighbors / Route-Maps
Adding R5-AS100 with routes 200.0.0.0/24. Set up direct connection between
R5 – R1. You will do eBGP on this link – R5@fa0/0 ↔ R1@fa0/0
J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
R5 - AS100
ebgp
FA0/0
FA0/0
Configs on R5 - AS100
interface Loopback200
description loopback200
ip address 200.0.0.1 255.255.255.0
!
interface FastEthernet0/0
description ebgp-R5-R1
ip address 100.0.0.1 255.255.255.252
!
router bgp 100
bgp log-neighbor-changes
neighbor 100.0.0.2 remote-as 99
!
address-family ipv4
neighbor 100.0.0.2 activate
neighbor 100.0.0.2 soft-reconfiguration inbound
no auto-summary
no synchronization
network 200.0.0.0
exit-address-family
!
!
!
ip forward-protocol nd
ip route 200.0.0.0 255.255.255.0 Null0 name BGP-PULL-UP
Originating Routes
E-BGP Peer
Note: On Junos
we write type
“internal or
external” &
peer-as
Create “routes” to
advertise. Remember, R5
is ONLY running BGP.
There isn't another
protocol to learn from and
inject into BGP
Configs on R1
router bgp 99
bgp log-neighbor-changes
neighbor 10.0.0.3 remote-as 99
neighbor 10.0.0.3 update-source Loopback0
neighbor 100.0.0.1 remote-as 100
!
address-family ipv4
neighbor 10.0.0.3 activate
neighbor 10.0.0.3 soft-reconfiguration inbound
neighbor 100.0.0.1 activate
neighbor 100.0.0.1 soft-reconfiguration inboundR5#show ip bgp neighbors 100.0.0.2 routes
BGP table version is 48, local router ID is 200.0.0.1
Network Next Hop Metric LocPrf
*> 10.0.0.0/24 100.0.0.2 0 99 i
*> 10.0.0.3/32 100.0.0.2 0 99 i
*> 10.0.0.11/32 100.0.0.2 0 99 i
*> 20.0.0.0/30 100.0.0. 0 99 i
*> 30.0.0.0/24 100.0.0.2 0 99 i
*> 100.0.0.4/30 100.0.0.2 0 99 i
Problem here: R1 is sending even /32s! Not good.
R1 needs a route-map OUT. See AS-99-OUT on
Cisco R1
neighbor 100.0.0.1 route-map AS-99-OUT out
!
ip prefix-list 10 seq 5 permit 10.0.0.0/24
ip prefix-list 30 seq 5 permit 30.0.0.0/24
ip prefix-list 40 seq 5 permit 40.0.0.0/24
!
!
!
route-map AS-99-OUT permit 10
match ip address prefix-list 10 30 40
set metric 600
R5#show ip bgp neighbors 100.0.0.2 routes
BGP table version is 56,
local router ID is 200.0.0.1
Network Next Hop Metric LocPrf
*> 10.0.0.0/24 100.0.0.2 600 0 99 i
*> 30.0.0.0/24 100.0.0.2 600 0 99 i
*> 40.0.0.0/24 100.0.0.2 600 0 99 I
New and better looking results
AS99 needs to be neat & tidy and
advertise only /24s. While it is OK to
have small /32s internally, it is NOT
OK to advertise such small blocks to
eBGP peers.
Another important route-map that R1
should include is to reject 0.0.0.0/0
from eBGP Peers.

6. 6. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco – How does R3 (Right) learn about 200.0.0.0/24
What we know so far. R5 – ebgp – R1. R1 ONLY learns 200.0.0.0/24 from R5. OK
How does R5 tell the LEFT network about new 200.0.0.0/24?
How does R5 tell the RIGHT network about new 200.0.0.0/24? J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
R5 - AS100
ebgp
FA0/0
FA0/0
Lets look at R2 – LEFT
R2#show bgp sum
BGP router identifier 10.0.0.11, local AS number 99
Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down
10.0.0.3 4 99 13 6 19 0 0 00:01:01 10
R2#show ip bgp neighbors 10.0.0.3 routes
BGP table version is 19, local router ID is 10.0.0.11
Network Next Hop Metric LocPrf Weight Path
*>i10.0.0.0/24 10.0.0 100 0 i
...
*>i200.0.0.0 100.0.0.1 0 100 0 100
R2#show ip route 200.0.0.1
Routing entry for 200.0.0.0/24
Known via "bgp 99", distance 200, metric 0
Tag 100, type internal
Last update from 100.0.0.1 00:02:38 ago
Routing Descriptor Blocks:
* 100.0.0.1, from 10.0.0.3, 00:02:38 ago
Route metric is 0, traffic share count is 1
AS Hops 1
Route tag 100
One bgp neighbor only
Many routes ..
focus on
200.0.0.0/24
Installed ROUTES
BGP 99
Default distance (AD)
“tag” = AS100
Demonstrates how RR works. There is no direct ibgp r'ship
between R2 & R1 with each other; but BGP routes still shared.
Also interesting is how R3 learns the 200.0.0.0/24 route since its ibgp
neighbor is 10.0.0.4 (J04) and not J03. J03 ↔ exchange routes ↔
J04, and J04 made it available to R3.
Delete “cluster 10.0.0.x” on J03 or J04, restart BGP and see how the
200.0.0.0/24 network disappears from the RIGHT network!
What we ALSO understand is that BECAUSE we have used a
100.0.0.0/30 IP Address on R1@fa0/0 it has been injected into our
OSPF table as internally used. The same also applies to 20.0.0.0/30
between J03 ↔ J04!
Cisco R3
R3#show bgp sum
BGP router identifier 10.0.0.12, local AS number 99
Neighbor V AS MsgRcvd MsgSent TblVer InQ
10.0.0.4 4 99 66 51 23 0 0 00:24:20
R3#show ip bgp neighbors 10.0.0.4 routes
BGP table version is 23, local router ID is 10.0.0.12
Network Next Hop Metric LocPrf Weight Path
*>i10.0.0.0/24 10.0.0.3 100 0 i
...
*>i200.0.0.0 100.0.0.1 0 100 0 100 i
R3#show ip route 200.0.0.1
Routing entry for 200.0.0.0/24
Known via "bgp 99", distance 200, metric 0
Tag 100, type internal
Last update from 100.0.0.1 00:25:27 ago
Routing Descriptor Blocks:
* 100.0.0.1, from 10.0.0.4,
Route metric is 0, traffic share count is 1
AS Hops 1
Route tag 100
J04 - RR
200.0.0.0 –
target DST route
100.0.0.1 –
target next-hop
Learned
from J04
Find 100.0.0.1
and you will find
200.0.0.0/24
R3 will use
another CPU
cycle to find
100.0.0.1/OSPF

7. 7. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco – Adding 2nd
EBGP Peers / Neighbors / Route-Maps
We are going to ADD another EBGP R6-AS200 to Cisco R2 and advertise 200.0.1.0/24
We will need to advertise the new AS200 to the LEFT & RIGHT
network as well as to AS100 which was established earlier. J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
R5 - AS100
ebgp
FA0/0
FA0/0
R6 - AS200
ebgp
FA0/0
FA0/0
Assuming that you have configured EBGP between AS99 (R2) ↔ AS200 (R6)
w/o any sort of route maps IN or OUT. You will expect to see that AS200
received many routes including AS100 prefix 200.0.0.0/24
R6 – AS200
R6#show ip route
B 200.0.0.0/24 [20/0] via 100.0.0.6, 00:02:38
20.0.0.0/30 is subnetted, 1 subnets
B 20.0.0.0 [20/0] via 100.0.0.6, 00:02:38
200.0.1.0/24 is variably subnetted, 2 subnets, 2 masks
C 200.0.1.1/32 is directly connected, Loopback200
S 200.0.1.0/24 is directly connected, Null0
10.0.0.0/8 is variably subnetted, 3 subnets, 2 masks
B 10.0.0.10/32 [20/0] via 100.0.0.6, 00:02:40
B 10.0.0.3/32 [20/0] via 100.0.0.6, 00:02:40
B 10.0.0.0/24 [20/0] via 100.0.0.6, 00:02:40
Confirm that
we can learn
AS100/prefix
R2 – AS99
R2#show ip bgp neighbors 100.0.0.5 routes
Network Next Hop Metric LocPrf Weight
*> 200.0.1.0 100.0.0.5 0 0 200 i
Confirm that
R02 is
receiving
routes from
AS200
The next QUESTION to ask. Is R5 (AS100) receiving
AS200/prefix (200.0.1.0/24)?
Quick answer is NO. This is 99% because we have a
ROUTE-MAP Out on R1(AS99) ↔ R5(AS100) and
there is an implicit DENY in the behavior.
Right now there are two remedial options:
a) modify ROUTE-MAP out on R1 ↔ R5 (AS100)
b) remove ROUTE-MAP out on R1 ↔ R5 (AS100)
Cisco R1 (AS99) → Cisco R5 (AS 100)
ip prefix-list 10 seq 5 permit 10.0.0.0/24
ip prefix-list 30 seq 5 permit 30.0.0.0/24
ip prefix-list 40 seq 5 permit 40.0.0.0/24
!
ip prefix-list AS200 seq 10 permit 200.0.1.0/24
!
route-map AS-99-OUT permit 10
match ip address prefix-list 10 30 40 AS200
set metric 600
* clear ip bgp 100 to reset bgp session
NEW
R5 – AS200 will now received AS200
R5#show ip route
100.0.0.0/30 is subnetted, 1 subnets
C 100.0.0.0 is directly connected, FastEthernet0/0
C 200.0.0.0/24 is directly connected, Loopback200
B 200.0.1.0/24 [20/600] via 100.0.0.2, 00:17:53
10.0.0.0/24 is subnetted, 1 subnets
B 10.0.0.0 [20/600] via 100.0.0.2, 00:17:53
30.0.0.0/24 is subnetted, 1 subnets
B 30.0.0.0 [20/600] via 100.0.0.2, 00:17:53
R5 receiving
AS200/prefix
NOTE:
The BIG problem here. What if we add AS300,
AS400, AS500 and so on? In order to ensure
that all routes are properly aggregated & all my
EBPG-Peers have visibility of each other, ALL
my ROUTE-MAP OUT will constantly require
modification/updates similar to R1..
How can this “inconvenience” be solved in a
more elegant way? Communities – Next Slide
Note: if J04 is set up properly, nothing extra
needs to be done for the RIGHT network to
learn AS100 & AS200

8. 8. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco – Using BGP Communities
What are BGP Communities? These are a special set of instructions that can be created in BGP
and re-advertised in BGP throughout the network. All routers destined to received the
re-advertised routes/prefix/ASN will have an option to receive
the community and perform a certain configured action.
In our set up what we are going to do is the following:
1. EBGP is already set up with AS100 & AS200
2. The PROBLEM is, if I add another EBGP AS300, I will
have a lot of work to do. I will need to go to every
router and reconfigure a AS-99-OUT to recognize
the new AS300/prefix, accept it and advertise it to
the participating / neighbor ASN... I am lazy.
3. See Slide 7 (above) where I had to add “ip prefix-list AS200”
on Cisco R1 to advertise 200.0.1.0/24 to EBGP
neighbor AS100. A similar ip prefix-list ASXXX would be
required on Cisco R2 to advertise 200.0.0.0/24 to
EBGP neighbor AS200.
4. The situation will become MORE complex if R6-AS200 has another
downstream customer... say, AS210. How would YOU configure
AS210 → AS200 → AS99 → TRANSIT → AS100 AND vice-versa?
Our BGP Community Example
Out example is plain and NOT extensive. But it is hoped that it will provide
a small insight into what we can use BGP Communities for and how it
can help automate the set up. We will ONLY consider the current setup
where we R2 has to learn about routes from R5-AS100 (200.0.0.0/24)
and re-advertise it to R6-AS200.
Our plan is to get J03 to learn AS100/prefix AND tag it with community 99:355
and tell R2 about it. R2 will receive 99:355, “install it” and re-advertise it to AS200.
J03 will ALSO have to learn AS200/prefix, tag it 99:355, tell R1. R1 will then have to receive it, install it & re-advertise it to AS100.
J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
R5 - AS100
ebgp
FA0/0
FA0/0
R6 - AS200
ebgp
FA0/0
FA0/0
E.G. AS300
The following MUST happen and we will use
BGP Communities to help us out.
1. J3 – Learns AS100/prefix (200.0.0.0/24)
2. J3 – Must tag it with 99:355
3. J3 – Must TELL Cisco R2
4. R2 - Must learn 99:355
5. R2 – Must ACCEPT it
6. R2 – Must TELL ebgp AS200
R6 – Will accept it anyway!
E.G. AS210

9. 9. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco – Communities – Auto learn, tag & announce
BGP Communities
Learn iBGP prefix & tag it with 99:355
J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
R5 - AS100
ebgp
FA0/0
FA0/0
R6 - AS200
ebgp
FA0/0
FA0/0
J03 – policy-option
policy-statement tag-355 {
term 1 {
from protocol bgp;
then {
community set ebgp-customers;
accept;
}
}
}
community ebgp-customers members 99:355;
Policy-Name
Learned from...
Add tags from community ...
We have created a policy-option named “tag-355”. It will learn
from protocol BGP and apply community name “ebgp-customers”:
99:355 as defined in community ebgp-customers a little lower
down the configs.
We need to APPLY this rule to the correct “protocols bgp group ...”
kjteoh@Junos-3> show conf.. proto.. bgp
group ibgp {
type internal;
local-address 10.0.0.3;
advertise-peer-as;
family inet {
unicast;
}
export [ ibgp_export tag-355 ];
cluster 10.0.0.3;
local-as 99;
neighbor 10.0.0.10 {
description Cisco-R1;
}
neighbor 10.0.0.11 {
description Cisco-R2;
}
Added – NEW
See Slide 3 for
a recap
After commit & clear
BGP neighbor .. test
your “new”
advertisements.
Now that this is done, what is J03 telling R2?
kjteoh@Junos-3> show route advertising-protocol bgp 10.0.0.11 200.0.0.0/24 detail
* 200.0.0.0/24 (1 entry, 1 announced)
BGP group ibgp type Internal
Nexthop: 100.0.0.1
MED: 0
Localpref: 100
AS path: [99] 100 I
Communities: 99:355
Cluster ID: 10.0.0.3
Originator ID: 10.0.0.10
J03 has learnt 200.0.0.0/24
and is tagging it with 99:355
AND is telling R2@10.0.0.11
R2 @10.0.0.11 needs to do now is create:
!
ip bgp-community new-format
ip community-list standard tag-355 permit 99:355
!
route-map AS-99-OUT permit 10
match ip address prefix-list 10 30 40 AS100
set metric 500
!
route-map AS-99-OUT permit 20
match community tag-355
set metric 355
Remove this.. see Slide 7
we don't want it anymore BUT
maintain prefix-list 10 30 & 40
Go to R6 and look at received
routes .. see metric 355
J3 learns
AS200
200.0.0.0
/24 untag
J3 tags
99:355
advertise it
R2 will learn
99:355 install
& advertise it
to AS200
Additional permit seq “20”

10. 10. Back to Basics – OSPF / iBGP (Route Reflectors) / eBGP Junos – Cisco - Notes
Disadvantage of using RR? What does bgp multipath do for us? Add redundancy to both
networks by building iBGP adjacency. Eg. J03 ↔ R3 & R4.
What would the situation be if R1 has a new eBGP customer
R7 AS-300? … BIG Question here!
Kjteoh 11/3/2016
J03-RR-0.3 J04-RR-0.4
20.0.0.0/30
30.0.0.X/24 40.0.0.0/24
R1-0.10
R2-0.11
R3-0.12
R4-0.13
AS99
left right
em0 em0
FA0/1
FA0/1 FA0/1
FA0/1
em1
R5 - AS100
ebgp
FA0/0
FA0/0
R6 - AS200
ebgp
FA0/0
FA0/0
ebgp
R7 - AS300