>Scenario:
>Let's say you have a class C subnet that you break off into two smaller
>networks:
>
>1.1.1.0/24 = 1.1.1.0/25 + 1.1.1.128/25
>
>Problem:
>You are announcing this through BGP, but your provider only allows you to
>advertise full class C addresses.  The problem lies in that you have to
>announce this network as 1.1.1.0/24 but still get traffic to 1.1.1.0/25 &
>1.1.1.128/25.
>
>Solution:
>So, there are two ways to do this:
>
>1) Use an aggregate-address with a suppress map to filter this.
>[snip]
>
>
>2) Use null routes
>[snip]



>Question:
>Which way do you all feel is better?
>
>I read in Boson's test that the preferred way is to use aggregate addresses,
>but this seems to be a much more complex way to do what a simple null route
>can accomplish?  Thought / comments?

First, let me deal with the problem definition. Unless you have a 
more advanced multihoming scenario, why would you ever want to 
advertise anything more specific than /24 to your provider?  The 
provider is going to use the same interface to send to both /25's, 
right? Telling the provider about the more specifics doesn't, in this 
example, give the provider any additional useful information.

Second, this is one of those cases where there's a CCIE Lab way and 
the right way.  It's my general impression that the proctors frown on 
solutions that involve static routes, and that they may explicitly 
say you can't use static routes. This appears to be meant to screen 
out candidates that use static routes when they can't figure out how 
to make dynamic routing do something it should (a Good Thing) and to 
demonstrate knowledge of the full IOS command set even when there is 
a neat and clean static route solution (a Bad Thing).

The blackhole null route, I would say, is the preferred technique of 
most real-world ISP routing engineers.  It doesn't appear to be a 
technique that Cisco emphasizes in classes or tests.  So Boson may be 
right in its approach to test preparation.  In my BGP papers at 
CertificationZone, I give both methods.

>

_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]

Reply via email to