cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
521
Views
3
Helpful
3
Replies

BGP redundancy

Antonio_1_2
Level 1
Level 1

Hi,

I have this situation now:


ROUTER R1 BGP peering toward ISP1
ROUTER R2 BGP peering toward ISP2


is there a way to manage that both

ROUTER R1 have BGP to ISP1 and ISP2
ROUTER R2 have BGP to ISP1 and ISP2
but not to change peer address from ISP perspective( ISP1 have still only one BGP peer toward me, and simmilar for ISP2)?

(i.e.I know HSRP will not work, but something simmilar to that)

Or the only way is to have 4 BGP peerings: R1-ISP1, R1-ISP2, R2-ISP1, R2-ISP2?

Thank you,
A

3 Replies 3

Tharak Abraham
Level 3
Level 3

Antonio,

When you ISP will not support multiple peering then you can rule out the option which you are telling.

For BGP to work, it needs config at both the ends directed at each other.

I dont think we can trick BGP as we do for other IGP's...-:)

Why do you want a design like this in  first place ?

We have other methods for redundancy rite ?

Any basic book on BGP will give you wonderful ideas on the same

Hi,

First up all In my view it is not a good practice.Even ya, there is a possiblity to configure full mesh BGP but How come an ISP1-PE will make ebgp neighborship with your side another ISP CE side.

If you want to make redundancy in your network, that is possible by configuring the below steps.

1.CE(ISP1) ----eBGP--->PE(ISP1)

2.CE(ISP1) ----iBGP(next-hop-self) --->CE(ISP2)  // Since We are configuring next-hop-self future in CE(ISP1),Even if the ISP2 link is down there will be no impact.The traffic flow will be CE(ISP2) --- CE(ISP1) -------eBGP--->PE(ISP1).

3.CE(ISP2) ----iBGP(next-hop-self) --->CE(ISP1).

You can make redundancy to youe network by configuring like this..

Regards

Karuppu

Hi,

in this way if link/port toward i.e ISP1 is down capacity will be reduced in half.

I just wanted to make sure that maybe there is some way to trick eBGP.

Thanks

A.

Review Cisco Networking for a $25 gift card