Re: Mitigating HTTP DDoS attacks?

2008-03-25 Thread Paul Wall
On 3/25/08, Peter Dambier [EMAIL PROTECTED] wrote:


 proc2pl might get you ideas, from the ISAON tools on


You know, for the last year or two I've heard you go on and on about IASON.
A few months ago I actually did download it and the only thing I can find in
it is an assortment of scripts to manage DNS zone files. I don't see
anything in there about auto detecting the network, automatically blocking
DDoS or any of the other artificial intelligence you purport it has.

Peter and Karin Dambier


I'm not sure how to interpret this. Are Peter  Karin the same person? You
be the judge.

http://wiki.piratenpartei.de/images/3/39/KarinPeter.jpg

Cesidian Root - Radice Cesidiana


Google searches on Cesidian Root reveal a rather scruffy man running this
alternative root out of Long Island, fighting for
secession from the United States. I'm sure Most Rev. Dr. Cesidio Tallini,
BS, PhD hc, CPC, RH-INHA, APP, AMBCS, MMPR, OEMTDV will get there, one day.

http://www.cesidianroot.com/


You might want to have the reverend doctor contact customer care, as the
website suggests. It seems the website is down.


Re: Qwest desires mesh to reduce unused standby capacity

2008-02-27 Thread Paul Wall
On Wed, Feb 27, 2008 at 9:37 PM, Frank Bulk - iNAME [EMAIL PROTECTED]
wrote:

 http://telephonyonline.com/access/news/ofc-qwest-optical-0226/
 To keep this OT as much as possible, my question is if a
 mesh-configuration
 of backup routes (where one link could provide 'protection' for many)
 would
 be considered a sufficient replacement for SONET rings, or if the Qwest
 CTO
 is really trying to get out of providing sub 50-msec protected loops and
 encouraging L3 and above protection schemes, so that they can even further
 over-subscribe their network.

 Frank


UU/MFS tried running IP on the 'protect' path of their SONET rings 10 years
ago. It didn't work then.

More seriously, you *can* avoid using protected links for IP (which is what
Qwest seems to suggest) easily, and allegedly using MPLS/FRR you could have
sub-second reroute times without having full dedicated protect path.

Building your network on preemptable links (the protect-side) as UU did back
in the day is probably of the I encourage my competitors to do this
solutions.

Paul Selling more grillz than George Foreman Wall


Re: YouTube IP Hijacking

2008-02-25 Thread Paul Wall

On Sun, 24 Feb 2008, Sargun Dhillon wrote:
 I don't know how large Pakistani Telecom is, but it I bet its not large
 enough that PCCW should be allowing it to advertise anything.

I think you're failing to take into account how multihoming generally
works.  The real fallacy here is that PCCW/BTN refuses to prefix-list
filter their customers, as evidenced by this and past leaks.  If
something productive can come from today's outage, it would be PCCW
beginning to do their part as responsible Internet citizens, given
(excuse the pun) peer pressure.

I'd also focus on the lessons learned from the un-official IP
Hijacking BOF held in San Jose, during which engineers and
researchers studied the extent to which obviously-bogus route
advertisements propagated across the public Internet.  At these
events,  prefixes such as 1/8 and 100/7 were advertised, and, by
Renesys/bgplay/route-views/etc data, accepted by 99% (?) of the
internet.  IP blocks that were hijacked before (like 146.20/16) were
announced with similar outcome.

Results were planned to be presented at the next NANOG, but they
shouldn't be a surprise to anyone in the industry: nobody filters.

Paul Wall