BGP Questions

Routing & Switching BGP
1

Question 1
AS 1 has a single border router with two peering links to ISP (AS2). How we can configure AS1 to ensure:

  1. Both links are utilized equally.
  2. One link as primary the other as backup link?
  3. One link is 1G, the other is 10G what needs to be done to make sure the 1G link doesn’t get saturated

Question 2

BGP peer calls and complains that their customers are complaining of a degraded performance with our services. What can be done in this situation?

Question 3

Can you please explain me the below concept? What it means. Can you please give an example?

Do not use the neighbor next-hop-self command to modify the next hop attribute for a route reflector when this feature is enabled for a route reflector client. Using the neighbor next-hop-self command on the route reflector will modify next hop attributes only for routes that are learned from eBGP peers and not the intended routes that are being reflected from the route reflector clients. To modify the next hop attribute when reflecting a route, use an outbound route map.

Its based on this link - http://www.cisco.com/c/en/us/td/docs/ios/12_2s/feature/guide/fs_bgpnh.html

Questions 4

A BGP goes to the Idle state due to an error. What happens next?

Questions 5

Do you any way to reduce inbound traffic via a given BGP neighbor/peer?

Question 6

Do you know any way to direct outbound traffic to a given BGP neighbor/peer?

Question 7
Does MED have significance when you have two connections to the same AS? How about different AS?

2

Hi Sahil,

  1. By default, BGP will only select one best path for each prefix. If your router is connected with two interfaces to one router on the other side then you could use ebgp multihop with loopback interfaces:

    If the router is connected to two routers, you’ll have to use some BGP attributes. On your end, you could configure local preference for some of the routes so that you can load balance:

    https://networklessons.com/bgp/how-to-configure-bgp-local-preference-attribute/

  2. This is probably not BGP related (unless you have problems with the BGP peering). This typically occurs with congestion...I would check for interface drops, interface load, etc. You might want to dive into QoS so that you can decide what traffic gets dropped in case of congestion.
  3. Hmm I'd have to lab this to confirm it but I think they are trying to say that you shouldn't use next hop self on the route reflector when you are already using next hop self on a route reflector client. If you would, the route reflector would overwrite the next hop that the route reflector client already has changed.
  4. BGP will normally keep retrying to establish a neighbor adjacency. Normally with "debug ip bgp" you can see the reason why it is failing.
  5. If you have multiple peerings with your neighbor AS, we typically use AS path prepending:

    If you only have one connection. You might want to use policing to limit your traffic:

    https://networklessons.com/quality-of-service/qos-traffic-policing-explained/

  6. This is best done with local preference: https://networklessons.com/bgp/how-to-configure-bgp-local-preference-attribute/
  7. MED is only compared when you have two neighbors in the same AS:

    However, there is a “Always compare MED” command that you can use to compare the MED from two different ASes. In real networks, it’s more common to use AS path prepending though.

Hope this helps!

Rene

3

Hi Rene

Just a question, regarding BGP, we generally use the neighbor {ip address} activate for what reason? From the research I have done, it exchanges information between BGP routers, however what is the difference when using
neighbor {ip address} which falls under router bgp AS Number
and
neighbor {ip address} activate which falls under address-family ipv4/vpnv4 of the router bgp AS number , does neighbor {ip address} not exchange BGP information? What is the purpose of the activate command?
Hope it makes sense as to what I am asking,
Thank you for your time

Kind Regards
Himal Bissessar

4

Himal,
The answer to your question has to do with what address families are automatically set to create neighbor relationships. By default, at least on Cisco, BGP will automatically create neighbor relationships in the IPv4 unicast address family. Because this is the most common one used, this is why you don’t often have to use the activate keyword. You can turn this behavior off, however via the following:
no bgp default ipv4-unicast

When you do run this command, simply typing in a neighbor {ip address} remote-as {ASN} would not be enough to form a neighbor relationship. You would need the additional activate command. By the way, turning off this default IPv4 behavior is sometimes useful in the situation where you want to use BGP for something besides IPv4 route sharing–like VPNV4 MPLS VPN routers, for example.