Hi,

You really didn't explain what issues you're experiencing in your test bed, but 
it sounds like you want simple iBGP peering. You'd do something like this on 
both routers.

/routing bgp peer
add instance=default remote-address=192.0.2.1 remote-as=36295 
update-source=loopback0

If you wanted to include R3 in your BGP config without having to build a full 
mesh between all routers then you want to make one router a Route Reflector 
(ie. R1).

## R1
/routing bgp instance set default client-to-client-reflection=yes

/routing bgp peer
add instance=default remote-address=192.0.2.2 remote-as=36295 
update-source=loopback0 route-reflect=yes

add instance=default remote-address=192.0.2.3 remote-as=36295 
update-source=loopback0 route-reflect=yes

## R2 & R3
/routing bgp peer
add instance=default remote-address=192.0.2.1 remote-as=36295 
update-source=loopback0

R1 will advertise routes received from R2 to R3, and vice versa thus 
eliminating the need for the configuration of a full mesh.

I'd like to know more about the issues you're having in your test bed as iBGP 
is usually a fairly straightforward setup.

--
Blake Covarrubias


On Aug 9, 2010, at 11:04 PM, Faisal Imtiaz wrote:

>> From what you have described, your probably want to run a BGP Route 
>> Refelctors... see page 50 of enclosed cisco doc.
> 
> I cannot tell you how to implement that on a mikrotik, but in these pages you 
> will see many scenarios that should give you the network design info you are 
> looking for.
> 
> Regards
> Faisal
> 
> ---------- Original Message ----------------------------------
> From: "Scott Carullo" <[email protected]>
> Reply-To: [email protected], WISPA General List <[email protected]>
> Date:  Mon, 9 Aug 2010 23:35:59 -0400
> 
>> I need some assistance...
>> 
>> I currently have two upstream connections at 100MB with a full BGP feed 
>> each into Mikrotik x86 routers wich are also running OSPF facing internally 
>> - no routes are distributed from BGP to OSPF.  The two BGP routers are 
>> located in two different cities and do not share BGP information between 
>> themselves - they are basically used to advertise out the networks to 
>> control the flow of traffic coming into our network.  Same upstream 
>> provider so they would have the same outgoing tables anyway.  We use OSPF 
>> path costs to direct outgoing traffic to the closest BGO router and 
>> basically the networks advertised follow the same paths in as out.  The 
>> /21s are advertised on both BGP routers with the /24s advertised to steer 
>> the inbound traffic through the same router the outbound traffic goes out.
>> 
>> Ok, that all works textbook style, no problems easy to manage for both OSPF 
>> (one internal area) and BGP. 
>> 
>> Now we need more bandwidth and have reached out to grab another provider 
>> with 100MB circuit in another city.  This is a different provider than the 
>> other two existing BGP routers are peered with and now there is a desire to 
>> have the BGP routers on my network exchange routes and select the best 
>> outgoing router based on the full BGP routing tables.  It's kind of hard to 
>> explain without the picture of the network to understand so here is a quick 
>> hand-drawn sketch of what the network layout looks like. Sorry its a mess 
>> just did it real quick while typing this email....  
>> http://www.brevardwireless.com/files/networkpic.pdf
>> 
>> I'm not really worried about router #3 participating in the BGP table 
>> exchanges because I will soon consolidate router #2 and #3 into the same 
>> router and combine bandwidth feeds into one fiber as well.  So for sake of 
>> discussion BGP router #3 does not exist, I drew it for completeness of my 
>> current situation.  I'm guessing I need to upgrade the tower routers on 
>> tower C and E to routers powerful enough to handle BGP and I have placed 
>> RB1100 on each so they are there.  Not interested in hearing about my 
>> router selections, just how to configure routing protocols to accomplish my 
>> goals (thanks).  We have attempted in testbed with 4 BGP routers in a row 
>> like they are on the network map with a full feed on each end and we are 
>> having a lot of luck.  Not sure if our testbed is fouled or it we just need 
>> more of a BGP clue.  Either way, some assistance would be greatly 
>> appreciated.
>> 
>> My preference is to not have BGP and OSPF exchange routes.  Since the BGP 
>> routers traverse the middle of my network I'll let OSPF do its magic of 
>> getting to the closest gateway and then let BGP direct it from there.  Once 
>> a packet hits any BGP router it will not have to cross a non-BGP router 
>> before it goes out in all cases.
>> 
>> I have considered professional consultation and will need to go there 
>> pending no internal progress on reaching our goals.  I'd rather ask first 
>> and learn myself and be able to manage what I have running without having 
>> outside 3rd party reliance.  Most of you probably understand that decision 
>> :)
>> 
>> Any way, if you have read this far and suffered through my network pic then 
>> thank you very much.  I appreciate your time and assistance. Hopefully this 
>> is an easy task for some of you.
>> 
>> Scott Carullo
>> 
>> Brevard Wireless
>> 321-205-1100 x102
>> 
>> 
>> 
>> 
>> 
>> 
> <bgp-toc.pdf>
> 
> --------------------------------------------------------------------------------
> WISPA Wants You! Join today!
> http://signup.wispa.org/
> --------------------------------------------------------------------------------
> 
> WISPA Wireless List: [email protected]
> 
> Subscribe/Unsubscribe:
> http://lists.wispa.org/mailman/listinfo/wireless
> 
> Archives: http://lists.wispa.org/pipermail/wireless/





--------------------------------------------------------------------------------
WISPA Wants You! Join today!
http://signup.wispa.org/
--------------------------------------------------------------------------------
 
WISPA Wireless List: [email protected]

Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless

Archives: http://lists.wispa.org/pipermail/wireless/

Reply via email to