Date: Tue, 14 Feb 2012 09:19:25 -0800
From: Gage Bystrom <themadichib0d@...il.com>
To: Lucas Fernando Amorim <lf.amorim@...oo.com.br>, 
	"full-disclosure@...ts.grok.org.uk" <full-disclosure@...ts.grok.org.uk>
Subject: Re: Arbitrary DDoS PoC

If the design is broken than the implementation is broken. Have you READ
your own source code? Do you understand what its actually doing? Rhetorical
questions of course but still.

Your poc calls curl multiple times via a list of proxies. No more, no less.
If you are going to claim that such a thing is an effective general
technique YOU have to back up that claim, not me or anyone else on this
list. I never bothered running it because anyone who read that simple
python code(which was a good thing its simple), can understand what it is
doing, and do a mental comparison to what they previously knew about the
subject of dos. Your poc does not demonstrate anything new, it demonstrates
existing knowledge that is generally known to not be an effective method
for dosing for all the reasons I explained in my previous mails.

I think its quite pedantic of you to only criticize me for calling out the
ineffectiveness of your poc. You did not address anything I or anyone else
said about your claim. If you think I am wrong or mistaken in my personal
assessment of your claim than you are the one who must show how and why to
defend your claim. Belittling someone who criticizes you is not
professional, not productive, does not give strength to your claim, and
does not make you right.

The end of the line is I don't care what you claim your code does, I care
about what the code does, and your code is not an effective general
technique for denial of service attacks.
On Feb 13, 2012 8:48 PM, "Lucas Fernando Amorim" <lf.amorim@...oo.com.br>
wrote:

> **
> I could argue that an attack targeted at a service, especially HTTP, is
> not measured by the band, but the requests, especially the heavier, could
> argue that a technique is the most inherent characteristic of multiple
> sources of traffic and still relying on trust. I could still say that is an
> implementation that relates only to say - Look, it exists!, I could still
> prolong explaining about overheads, and using about the same time many
> sites that make the requests, thus reducing the wake of a failure, even if
> you say easily diagnosable.
>
> But I'd rather say that it is actually very pedantic of you label
> something as inefficient, especially when not done a single test, only the
> pedantic observation of someone whose interests it is reprehensible. I will
> not say you're one of those, but this is really an attitude typical of this
> kind, which is certainly not a hacker. Thanks to people like that, do not
> know if you like, there are many flaws yet to be explored.
>
> If anyone wants more information, obviously I will ask to send an email or
> call me to give a presentation, I will not think about anything. My goal in
> was invited researchers to study DDoS on this model, because anytime
> someone can direct thousands to generate a network congestion.
>
> On 13-02-2012 11:17, Gage Bystrom wrote:
>
> Uhh...looks pretty standard boss. You aren't going to DoS a halfway decent
> server with that using a single box. Sending your request through multiple
> proxies does not magically increase the resource usage of the target, its
> still your output power vs their input pipe. Sure it gives a slight boost
> in anonymity and obfuscation but does not actually increase effectiveness.
> It would even decrease effectiveness because you bear the burden of having
> to send to a proxy, giving them ample time to recover from a given request.
>
> Even if you look at it as a tactic to bypass blacklisting, you still
> aren't going to overwhelm the server. That means you need more pawns to do
> your bidding. This creates a bit of a problem however as then all your
> slaves are running through a limited selection of proxies, reducing the
> amount of threats the server needs to blacklist. The circumvention is quite
> obvious, which is to not utilize proxies for the pawns....and rely on shear
> numbers and/or superior resource exhaustion methods....
> On Feb 13, 2012 4:37 AM, "Lucas Fernando Amorim" <lf.amorim@...oo.com.br>
> wrote:
>
>> With the recent wave of DDoS, a concern that was not taken is the model
>> where the zombies were not compromised by a Trojan. In the standard
>> modeling of DDoS attack, the machines are purchased, usually in a VPS,
>> or are obtained through Trojans, thus forming a botnet. But the
>> arbitrary shape doesn't need acquire a collection of computers.
>> Programs, servers and protocols are used to arbitrarily make requests on
>> the target. P2P programs are especially vulnerable, DNS, internet
>> proxies, and many sites that make requests of user like Facebook or W3C,
>> also are.
>>
>> Precisely I made a proof-of-concept script of 60 lines hitting most of
>> HTTP servers on the Internet, even if they have protections likely
>> mod_security, mod_evasive. This can be found on this link [1] at GitHub.
>> The solution of the problem depends only on the reformulation of
>> protocols and limitations on the number of concurrent requests and
>> totals by proxies and programs for a given site, when exceeded returning
>> a cached copy of the last request.
>>
>> [1] https://github.com/lfamorim/barrelroll
>>
>> Cheers,
>> Lucas Fernando Amorim
>> http://twitter.com/lfamorim
>>
>> _______________________________________________
>> Full-Disclosure - We believe in it.
>> Charter: http://lists.grok.org.uk/full-disclosure-charter.html
>> Hosted and sponsored by Secunia - http://secunia.com/
>>
>
>

Content of type "text/html" skipped

_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.grok.org.uk/full-disclosure-charter.html
Hosted and sponsored by Secunia - http://secunia.com/

Powered by blists - more mailing lists