[Xorp-users] Click configuration generator on XORP

Rafael Paoliello Guimaraes rafael.guimaraes@ac.upc.edu
Tue, 25 Jan 2005 15:34:08 +0100


Hi Eddie,

I know that this is not a bug, for me it was just redundant. I was just 
trying to understand why it was done this way, maybe I was missing 
something. In fact, I thought that maybe this should be interesting in 
the case that we had more than one interface connected to the same 
subnet. But this was the only case I could think of where this 
configuration may be more efficient. Anyway, thank you for your time...

Cheers,

===========================================
  Rafael Paoliello Guimaraes
  PhD Student - Computer Networking Group
  Department of Computer Architecture (DAC)
  Polytechnic University of Catalonia (UPC)
  Phone: +34-934017187   Fax: +34-934017055
  URL: http://people.ac.upc.es/rpaoliel
===========================================


Eddie Kohler wrote:
> Hi Rafael,
> 
> I think you're right, that this is not necessary, but it's the way we've 
> written Click IP configurations since we wrote Click.  (Which is like 
> more than 5 years ago.  Shit, time flies.)  ARP replies received on 
> interface X could be sent only to the ARPQuerier for interface X, and 
> the host.
> 
> The current design makes sense if an ARP reply might arrive on an 
> interface different from the interface that sent the request.
> 
> However, the current design isn't exactly a bug: it behaves correctly.
> 
> Eddie
> 
> 
> On Jan 21, 2005, at 11:06 AM, Rafael Paoliello Guimaraes wrote:
> 
>> Hello,
>>
>> I was taking a look at the configuration file generated by XORP to be 
>> used in Click and there is a point that I didn't really understand.
>>
>> Whenever an interface receives an ARP Reply, makes n+1 copies of it 
>> (through the Tee element) and send one copy for each of the n 
>> interfaces (in fact to their ARPQuerier element) and 1 copy to the 
>> host (in the case that click is running in user-level, this last ARP 
>> reply is discarded).
>>
>> Why does an ARP reply need to be sent to the ARPQuerier of other 
>> interfaces? Is there any case where a packet may be waiting an ARP 
>> reply to be sent through an interface and this ARP reply comes from 
>> another interface? Am I missing something?
>>
>> Best regards,
>>
>> -- 
>>
>> ===========================================
>>  Rafael Paoliello Guimaraes
>>  PhD Student - Computer Networking Group
>>  Department of Computer Architecture (DAC)
>>  Polytechnic University of Catalonia (UPC)
>>  Phone: +34-934017187   Fax: +34-934017055
>>  URL: http://people.ac.upc.es/rpaoliel
>> ===========================================
>> _______________________________________________
>> Xorp-users mailing list
>> Xorp-users@xorp.org
>> http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-users
> 
> 
> _______________________________________________
> Xorp-users mailing list
> Xorp-users@xorp.org
> http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-users
>