On Sat, 1 Jul 2006, Saikat Guha wrote:
Date: Sat, 01 Jul 2006 14:27:59 +0100
From: Saikat Guha <[EMAIL PROTECTED]>
Subject: Re: [p2p-hackers] Percentage of NAT-to-NAT pairs
On Sat, 2006-07-01 at 03:21 -0400, Michael J Freedman wrote:
http://illuminati.coralcdn.org/stats/
Very interesting!
The one I found particularly worrying is the Use of Reserved Addresses
behind NATs
(http://illuminati.coralcdn.org/imgs/datagen/nats-by-prefix-reserved-bar.png)
Do you have any stats on whether some NAT'ed subnets are
(inappropriately) using non-reserved blocks. For example: 41.0.0.0/8
used to be reserved and was being used internally by FastWeb (Italy,
Turin); it then got assigned to AfriNIC
(http://www.merit.edu/mail.archives/nanog/msg00288.html) Consequently,
some people in Turin cannot reach some people in Africa -- NAT or no
NAT.
Hi Saikat,
We don't really have such information. The problem you bring up is more
an issue of legal route advertisements and is somewhat orthogonal to NATs.
I assume from the above example that FastWeb probably isn't announcing
41.0.0.0/8 anymore given its re-allocation; if that is the case, it's
unclear to me how one could really differentiate between the two (when
coming from NATs) without a priori knowledge.
I guess one way to try to detect such behavior is if the *clients* can
perform traceroutes and then we try to map their egress routers based on
geo and topological information. Unfortunately, trying to traceroute
public IPs in server->client direction---which we *can* do---probably
won't work, given that our packets will correctly just go to the African
subnet, modulo route announcement conflicts.
Anyway, we don't currently have such information.
--mike
-----
www.michaelfreedman.org www.coralcdn.org
_______________________________________________
p2p-hackers mailing list
[email protected]
http://lists.zooko.com/mailman/listinfo/p2p-hackers
