PeraHoman
asked on
cases of bgp being redistributed into igp
why would one do this? say we have two isp per/cer, two ebgp peers, ibgp peering between cer, and to mpls bsckbone (no internet routes). igp is ospf. nothing fancy, just two circuits for redundancy to ensure network is up.
is it for faster reconvergence if one isp goes down through ospf?
with both CER being ibgp neighbors, is that enough for both routers to know each others bgp recieved routes?
what if there us no redistribute, how would the routing table look?
is it for faster reconvergence if one isp goes down through ospf?
with both CER being ibgp neighbors, is that enough for both routers to know each others bgp recieved routes?
what if there us no redistribute, how would the routing table look?
ASKER
Reference image
Note030816.pdf
So this is how a lot of my sites are set up.
ATT and Verizon routers are PERs.
R1
It has eBGP peer to ATT. We configure static BGP network prefixes which advertises only the LAN networks downstream to ATT. It has iBGP peer with R2. On R1, we run OSPF P2P on /30 downstream to Core 1 (sorry not labelled on image), and OSPF P2P between cores (not labelled). We redistribute BGP into OSPF (85%-95% of BGP received routes via ISP are 10.0.0.0/8 networks in our MPLS backbone with the rest being networks of other businesses). Also Core 1 has default route to R2. Typical route-map filters added to prevent transit and improper routing, but not the main question.
R2
Mirror image, but to Verizon. eBGP peer to Verizon, with static BGP network prefixes from LAN being advertised. iBGP peer to R1, OSPF P2P network between /30 of R2 and Core 2. OSPF P2P network between Core 1 and Core 2. Also Core 2 has default route to R1.
My main question is why is BGP redistributed into OSPF in R1 as well in R2? Shouldn't the iBGP peering between R1 and R2 be enough for them to know where to go? Plus there's the default route from Core 1 to R1, and Core 2 to R2 via static default AND OSPF default-information originate.
I have theories, but need to some insight.
Note030816.pdf
So this is how a lot of my sites are set up.
ATT and Verizon routers are PERs.
R1
It has eBGP peer to ATT. We configure static BGP network prefixes which advertises only the LAN networks downstream to ATT. It has iBGP peer with R2. On R1, we run OSPF P2P on /30 downstream to Core 1 (sorry not labelled on image), and OSPF P2P between cores (not labelled). We redistribute BGP into OSPF (85%-95% of BGP received routes via ISP are 10.0.0.0/8 networks in our MPLS backbone with the rest being networks of other businesses). Also Core 1 has default route to R2. Typical route-map filters added to prevent transit and improper routing, but not the main question.
R2
Mirror image, but to Verizon. eBGP peer to Verizon, with static BGP network prefixes from LAN being advertised. iBGP peer to R1, OSPF P2P network between /30 of R2 and Core 2. OSPF P2P network between Core 1 and Core 2. Also Core 2 has default route to R1.
My main question is why is BGP redistributed into OSPF in R1 as well in R2? Shouldn't the iBGP peering between R1 and R2 be enough for them to know where to go? Plus there's the default route from Core 1 to R1, and Core 2 to R2 via static default AND OSPF default-information originate.
I have theories, but need to some insight.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Simply - it is too many routers for OSPF area, whole network would have long outages.
To have best exit point for your traffic. If routers run iBGP between the border routers in that case routers can find the best exit point how to reach some external network.