High Level BGP question on network statements

I'm reading up on BGP and I think I have the jist of it. The only thing I'm not 100% on since I haven't worked with it in the field is the network statements. When advertising networks through BGP I assume you have to enter a network statement for each network so if there are 20 subnets then I have to have a network statement for each one since redistributing an IGP into BGP is a big no-no.

Can I use a summary address as a network statement? I'm thinking it'll be real cumbersome plugging away so many network statements if its a large network.

Thanks

Find more posts tagged with

Save $250 on 2025 certification boot camps from Infosec!

Book now with code EOY2025

Button

Comments

d4nz1g

I haven't touched BGP for a while (1 to 2 years). Last time I needed to advertise a summary, I used these commands:

network 10.0.0.0 255.0.0.0
ip route 10.0.0.0 255.0.0.0 null0

These two commands will: put the route to 10.0.0.0/8 in your routing table, and advertise the network to your peers.

networker050184

I think what you are looking for is the aggregate-address command.

late_collision

techiie wrote: »

redistributing an IGP into BGP is a big no-no.

The short answer is, it depends. I also think you have it backwards, redistributing BGP into an IGP is the big no-no. The big no-no mantra stems from the idea of dumping a large set of BGP routes, such as the global internet routing table, into an IGP. This is one of those foolish mistakes that you would only have to make once, and you'd probably have to update your resume too.

There are designs in which an IGP and BGP interact and redistribute between each other. A design might run BGP in the core of your network, while each distribution block runs its own IGP. The purpose might be different administrative domains, pools of mixed vendor gear, or maybe even legacy gear that supports only RIP. Typically summarization, filtering, and default routing is employed, but it is an example that gives your routing scheme a hierarchy and illustrates how an IGP and BGP can work together.

techiie wrote: »

Can I use a summary address as a network statement? I'm thinking it'll be real cumbersome plugging away so many network statements if its a large network.

You could, but I think networker050184 is steering you the correct direction with the aggregate-address command.

techiie

Thanks guys for the input! So how is it commonly done from and enterprise HQ perspective to ISP assuming its most likely multi-homed? Do you guys just redistribute the IGP into BGP and call it a day? I imagine depending on the scope of the enterprise advertising the networks with the network command could get a little crazy?

networker050184

In a simple set up like that just do as d4nz1g pointed out. Usually as an enterprise you are going to have only a couple blocks you advertise to the internet.

techiie

Ok this is when real-world experience is a major difference for understanding something in comparison to self study. I'm labbing all sorts of topics on BGP with whatever ip addressing scheme I think of and I just realized I'm missing something.

I was driving home and was thinking about the entire process and started to think where would I apply NAT and all that good stuff and I was quickly confusing myself.

Please Correct me if I am wrong:
The advertised network(s) from the iBGP edge router within your own company AS is the public IP address given by the ISP right not the IGP network running within your network right? I think in my previous posts thats what was throwing me off.

In that setup would NAT be configured on that router or a downstream core router/l3 switch?

ISP AS100->Your companys BGP router AS 200 -> 6506e chassis

networker050184

It really depends on the setup, but usually if you're running BGP you'll have an edge router that peers with one or more ISPs. Behind this will be your network infrastructure. Whether that has public or private IPs, NAT, etc. depends on a lot of factors. Usually it's going to be a FW doing your NAT though.

techiie

Thank you for answering my question on NAT.

Ok, so I take it the main reason why you would redistribute an IGP into BGP if BGP was running within the CORE. So it would be used to as a "Hub" for different routing domains to communicate with each other. Which would be the example late_collison mentioned

Besides that situation only the block of public ip addresses will get advertised from an internet edge router (BGP speaker) to other ISP's

Right?

networker050184

Well, there really aren't any reasons I can think of off the top of my head to redistribute an IGP into BGP. Not good reasons that aren't better solved in other ways anyway.