Manual:Simple BGP Multihoming: Difference between revisions
No edit summary |
No edit summary |
||
Line 126: | Line 126: | ||
</pre> | </pre> | ||
Configuration above is only for packets going to our network. There are several options how to deal with packets going from our network: | |||
* leave gateways as in [[#Main/Backup link setup | main/backup]] configuration - this will result in only one link utilized and asymmetric routing | |||
* use [[PCC#Policy_routing | policy routing]] to force outgoing packets over the same link as incoming | |||
* use BGP to receive full routing tables from both peers and using BGP attributes make part of the routes available through one link and other part through another link. For example, traffic local to your country is sent over ISP1 the rest is sent over ISP2. | |||
All those methods are covered in other articles and will not be shown here. | |||
Latest revision as of 13:52, 5 November 2010
Applies to RouterOS: all
Setup
Ilustration below shows simple multihomed BGP setup. This setup can be used for load sharing between ISPs or one ISP as main and other ISP as backup link.
Lets say that local Internet registry assigned to us two /24 networks: 10.1.1.0/24 and 10.1.2.0/24 and our AS is 30 (Private AS cannot be used in such setups). First network entirely is used for workstations in our corporate network. Part of the other network is also used for workstation and another part is reserved for server. At this point our company has only one server with address 10.1.2.130
The goal is advertise our assigned networks to BGP peers and use only one provider as main link, ISP2 link is for backup only.
Note: This example does not show how to provide connectivity between core router, local networks and servers
BGP Peering
Consider that IP connectivity between ISPs edge routers and Our Core router is already set up and working properly. So we can start to establish BGP peering to both ISPs.
#set our AS number /routing bgp instance set default as=30 #add BGP peers /routing bgp peer add name=toISP1 remote-address=192.168.1.1 remote-as=10 add name=toISP2 remote-address=192.168.2.1 remote-as=20
If everything is set up properly, peer should have E (established) flag and router should receive bunch of BGP routes from both ISPs
[admin@RB1100test] /routing bgp peer> print Flags: X - disabled, E - established # INSTANCE REMOTE-ADDRESS REMOTE-AS 0 E default 192.168.1.1 10 1 E default 192.168.1.2 20
Network Advertisements and Routing Filters
Now we can start to advertise our networks and filter out all other unnecessary advertisements.
First step is to advertise our networks
/routing bgp network add network=10.1.1.0/24 synchronize=no add network=10.1.2.0/24 synchronize=no
Next step is to specify which routing filter chains will be used
/routing bgp peer set isp1 in-filter=isp1-in out-filter=isp1-out set isp2 in-filter=isp2-in out-filter=isp2-out
in-filter is for incoming (received) prefixes, out-filter is for advertised prefixes.
Main/Backup link setup
After chains are specified we can accept our networks and drop everything else as we are not transit provider. As you know one of the BGP attributes that influence best path selection is AS Path length (shorter AS Path is more preferred). So as we want ISP2 to be backup only, we will use BGP AS prepend (increase length of AS path) to force incoming traffic through ISP1.
Outgoing filters to ISP1:
/routing filter #accept our networks add chain=isp1-out prefix=10.1.1.0/24 action=accept add chain=isp1-out prefix=10.1.2.0/24 action=accept #discard the rest add chain=isp1-out action=discard
Outgoing filters to ISP2:
/routing filter #accept our networks and prepend AS path three times add chain=isp2-out prefix=10.1.1.0/24 action=accept set-bgp-prepend=3 add chain=isp2-out prefix=10.1.2.0/24 action=accept set-bgp-prepend=3 #discard the rest add chain=isp2-out action=discard
We also do not need any routes from both ISPs, because default route is used to force outgoing traffic through ISP1 and leave ISP2 as backup.
/routing filter add chain=isp1-in action=discard add chain=isp2-in action=discard
/ip route add gateway=192.168.1.1 check-gateway=ping add gateway=192.168.2.1 distance=30 check-gateway=ping
Load sharing setup
Using previous setup we are kind of wasting one link. So it is possible to redesign our setup as illustrated below to utilize both links.
The same as in previous setup BGP AS prepend will be used to achieve our goal. This time we will advertise one of the netowrks to ISP1 without prepend and another network prepended three times. The opposite for ISP2.
Outgoing filters to ISP1:
/routing filter #accept our networks and prepend second network add chain=isp1-out prefix=10.1.1.0/24 action=accept add chain=isp1-out prefix=10.1.2.0/24 action=accept set-bgp-prepend=3 #discard the rest add chain=isp1-out action=discard
Outgoing filters to ISP2:
/routing filter #accept our networks and prepend first network add chain=isp2-out prefix=10.1.1.0/24 action=accept set-bgp-prepend=3 add chain=isp2-out prefix=10.1.2.0/24 action=accept #discard the rest add chain=isp2-out action=discard
Configuration above is only for packets going to our network. There are several options how to deal with packets going from our network:
- leave gateways as in main/backup configuration - this will result in only one link utilized and asymmetric routing
- use policy routing to force outgoing packets over the same link as incoming
- use BGP to receive full routing tables from both peers and using BGP attributes make part of the routes available through one link and other part through another link. For example, traffic local to your country is sent over ISP1 the rest is sent over ISP2.
All those methods are covered in other articles and will not be shown here.
[ Top | Back to Content ]