North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: Cable Modem [really responsible engineering]

  • From: Miquel van Smoorenburg
  • Date: Wed Jun 27 07:37:52 2001
  • Distribution: cistron
  • Newsgroups: lists.nanog

In article <[email protected]>,
Chris Adams  <[email protected]> wrote:
>Once upon a time, Miquel van Smoorenburg <[email protected]> said:
>> When the BRAS requests config info when the circuit goes up (using
>> radius) or when it acts as a DHCP relay, it includes the VPI/VCI
>> of the ATM channel in the request. That means that you can assign
>> IP addresses based on the physical connection rather than the MAC
>> address, and this is what we do [well, will do soon anyway ;)]
>
>Okay, but how do you keep the end user from putting a different IP in
>their computer?

The BRAS equipment we use, redback SMSes, can filter out IP addresses
with invalid source addresses. Like cisco's ip verify unicast reverse-path

>Also, how do you prevent the user from trying to forge someone else's
>IP address or even MAC address in outgoing packets?

Like I said, the SMSes we use filter IP, and it doesn't use real
bridging even within the same subnet, it does proxy arp. So if a
customer arps for another IP in the same subnet, the SMS will answer
the ARP request itself, it will not be bridged.

Unfortunately I have not been able to play with Cisco's 6400
series yet to see if they offer the same functionality - not that
we're not happy with our current equipment but I'd like to know
a bit more about how other equipment behaves. However from the
docs I get the impression that Cisco calls this IRB.

>Without protecting
>against forged packets, I don't see how to provide accountability when
>someone attacks.

Very true. The BRAS must be able to protect from IP spoofing and
it must do proxy arp instead of real bridging.

Mike.