REdistribute BGP to OSPF

My WAN connection is going through a private MPLS network. My layer 3 core switch is BGP peering with my ISP CE router. My core is also connecting to a FW for the backup internet. I am using 2 floating static default routes with higher AD route next-hop to the FW.

Now I am in the process of adding a router to the setup. So the router will take the function of the WAN router. BGP peering from the core will be moving to the WAN router. I am thinking of configuring OSPF between my WAN router and my core switch. So my questions are:
- Should I redistribute BGP into OSPF on the WAN router? My WAN is using MPLS VPN so I only see the routes from my remote sites via BGP.
- If redistribution is necessary, will my failover with the floating static default routes still work?
- If I don't redistribute, how will this whole thing work? Will it work?
Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

I would suggest that the size of the network does not justify the additional complexity of multiple routing protocols.
leblancAccountingAuthor Commented:
I do not have a choice as far as the BGP is concerned on the WAN side. For my internal network, instead of doing static routes, I was thinking dynamic routing for ease of management.
BGP as your EGP and OSPF as your IGP are standard design practices.

Extenal routes are known via BGP and Internal routes are know via OSPF, again standard. I would not redistribute BGP into OSPF, this could cause you many problems down the line. Your BGP speaker speaks OSPF so why not generate a default route into the OSPF domain so all traffic destined to routes not in the OSPF routing table will route to the edge where BGP routes can take over routing to the destination. Make sense?

harbor235 ;-}

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
Big Business Goals? Which KPIs Will Help You

The most successful MSPs rely on metrics – known as key performance indicators (KPIs) – for making informed decisions that help their businesses thrive, rather than just survive. This eBook provides an overview of the most important KPIs used by top MSPs.

leblancAccountingAuthor Commented:
I was planning to have a default route pointed to the MPLS CE router from my WAN router. Also, on the WAN router, configure BGP on the WAN link interface and OSPF with default-information originate on the internal interface. the default-information statement is to advertise my default route to my OSPF neighbor.
Yeah, his WAN is internal routing, although it's going to look like eBGP because the routes are coming from his MPLS provider.

The difficulty arises when you have multiple MPLS circuits, each with a different AS number, sometimes from different providers.  This can get tricky when you are advertising your internal addresses through the MPLS mesh.

I always had good luck with GRE tunnels.  GRE tunnels make an MPLS circuit look like a point-to-point connection, and can be considered a single hop from your internal routing protocol's perspective.  Then you can just use one routing protocol for all of your internal segments, and the MPLS/BGP configurations are only locally significant to the devices handling each WAN link.
Without seeing your architecture i am guessing here,

Do you have a diagram of your setup? stick drawing would be fine.

harbor235 ;-}
Routes available via MPLS should be routes assigned to your domain, more specifically address space that you are registered for.  So here it is easy to summarize and route to destinations via the MPLS path.

Internet routes are all other public routes, should be very easy to craft route logic to route to the proper destinations via the preferred path.

harbor235 ;}
leblancAccountingAuthor Commented:
Plz see attached.

Who are you dealing with that they require you to register your internal addresses?  

MPLS should be completely private, and the provider has no need to understand your address space or topology.  The provider can and probably does have other customers that overlap your address space, but neither you nor they should care.
I see now, why not do the following:

2911 CE2 advertises prefixes for ASN 30 and 40 to MPLS cloud
2911 CE1 advertises prefixes for ASN 10 and 20 (DIO would do this too)
2621 advertises BGP DIO, also here is where I would advertise OSPF into BGP
2911 FW advertises DIO into OSPF

Get the picture, hopefully I did not miss something. Would be allot easier if you connected all ASNs to the MPLS cloud. Are you paying for private peering connections between sites as well?

MPLS can be private if you like, that does not preclude you from using assigned address space. In fact most customers use assigned address space , many organizations like to avoid NAT if possible.

The provider carriers all customer routes so its pretty easy to see customer networks. I agree that with MPLS you can have IP address range overlap, but if you have IP space and you do not want to NAT why go through the pain.  Many organizations use RFC1918 everywhere and with little planning, in many cases it becomes difficult to use. Just my thoughts.

harbor235 ;}
leblancAccountingAuthor Commented:
I do not need to redistribute OSPF into BGP. My question was around the BGP to OSPF redistribution at the 2911 WAN routers. It seems not necessary to me. That is why I ask if this is typical for a private MPLS to redistribute BGP into OSPF.

The MPLS is private. They advertise our private IP subnets within their BGP.
According to your diagram I do not see where your WAN 2911s are members of any OSPF routing domain. I assume you use BGP to exchange routes with your MPLS provider? I assume 2911 WAN routers to be the 2911 MPLS routers?

It is more common to redistribute iBGP routes into an IGP, however, If you do not have many eBGP routes then it would be fine to redistribute BGP into OSPF. Redistributing routes into OSPF will assign all routes as external type 2 routes which in case of route decision making is less preferred than non external type routes.

In your diagram the OSPF speakers are 2911FW, 3560,2621, 2911 WAN-remote, 3560.

Sounds like you have a good grasp whats going on, your architecture is single threaded so you should be fine.

harbor235 ;}
leblancAccountingAuthor Commented:
Yes. The 2911 MPLS routers are the provider routers. We are EBGP between their 2911 and my WAN 2911. I just don't see any advantage to redistribute BGP to OSPF on my WAN 2911 router. Unless you guys see the advantages. Thanks.
The advantage would be failover to less preferred routes in an HA environment, which you do not have.
If you do not redistribute then you will need to use default-information-originate, nexthop-self and other techniques to achieve end to end traffic flows.

harbor235 ;}
leblancAccountingAuthor Commented:
Let me try.
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Networking Protocols

From novice to tech pro — start learning today.