On Wed, 2006-07-05 at 11:36 -0700, Martin Casado wrote:
> I guess the problem is fundamental, if we are using some 
> public prefix (say
> 1.2.3/24) and I send a packet to 1.2.3.11, the network cannot determine 
> whether I'm intending
> to send traffic internally or externally.  

Correct. Since the destination address (global) happens to match the
local subnet, the host will simply issue and ARP for the destination
(and either get a response from the wrong guy, or not get any response).
In any case, the host will not even attempt to forward to its next-hop
router/NAT (which is contacts for destinations not on the local subnet).

> This is a pretty serious blunder ... though may be difficult to 
> determine remotely due to the interposition of
> proxies and tunnelling.

Do you have a feel for how often corporations (big enough to have their
own address block) are subjected to their ISP's proxy?

I know lots of ISPs in Europe proxy dial-up/DSL/Cable users (who don't
own an address block obviously); I don't know how common it is to proxy
an address-block-owning entity. If it is not very common, then checking
the owner for the address-block for the client, and the owner for the
address-block for the proxy/NAT would shed some light on the matter.

On Wed, 2006-07-05 at 10:53 -0700, Martin Casado wrote:
> btw, do you know FastWeb's public address block? We can poke through
> our data and see if we have anything on it.

FastWeb is AS12874; address block details:
http://www.cidr-report.org/cgi-bin/as-report?as=AS12874

cheers,
-- 
Saikat

Attachment: signature.asc
Description: This is a digitally signed message part

_______________________________________________
p2p-hackers mailing list
[email protected]
http://lists.zooko.com/mailman/listinfo/p2p-hackers

Reply via email to