VoIP pentest ?

[Mark Teicher <mht3 () earthlink net>]

 There are several ways to provide some information to your question.  As in my previous response, there are some 
commercial products available to assist in VOIP Network readiness <QOS, Jitter, etc>.  The other is conducting the 
audit itself.

1. Physical Access - A sledge hammer or unplugging the pbx can do more damage than finding a X-Windows exploit.  <good 
to insert into any report, "We found X running on port 6000, and the janitor has the ability to unplug the PBX to plug 
in their heavy-duty floor buffer">  <Which would be given more weight in this case??">

2. System Configuration - a majority of TDM PBX's that have the ability to support a MEDIA/IP Card usually have a 
support modem attached to it <for vendor/customer support>, most network scanners will not pick the support modems up, 
a majority if times, an organization often forgets the number or forgets it is even there.  Better, the vendor 
providing support can't provide logs for when they log in.

3. Log Analysis - not very useful especially when examining legacy PBX network equipment, good luck looking for source 
and dest ip, correlating SMDR output to the Call Detail Recording System in place, another item to look is at the 
maintenance support companies. 
<How many times did you see calls over the PSTN to the PBX manufacturer more than x number of minutes, due to Cesar the 
support tech got caught up in a conference call and forget to close the PBX Remote Console ringing up minutes of your 
long distance?>  <Again, not covered by your <insert traditional network security scanning tool>

4. Disaster Recovery Plan - does the organization have a backup plan?? How much money is lost when a network outage 
occurs, how much money does it cost when the telecommunication network goes out, multiply that by number of dollars to 
migrate to VOIP and then include traditional network outage costs into the new figure.  



-----Original Message-----
From: Jerry Shenk [mailto:pen-test () securityfocus com] 
Sent: Thursday, December 09, 2004 10:57 AM
To: 'Mark Teicher'; pen-test () securityfocus com
Subject: RE: VoIP pentest ?

So, Mark - what are some of the good tools for testing a network for VOIP readiness?  I've got a local company that is 
"real hot" on VOIP....like it's gonna be the end-all to every problem.  I suppose it can help a few issues but they 
need a little help giving a little thought to some of the performance and security issues.

-----Original Message-----
From: Mark Teicher [mailto:mht3 () earthlink net]
Sent: Monday, December 06, 2004 9:28 PM
To: pen-test () securityfocus com
Subject: Re: VoIP pentest ?


Actually, the question for VOIP pen-testing should be split into two
issues:

1. How many vulnerable is a network with VOIP ?
2. Is the network ready for VOIP?
3. VOIP Attack suite

1. Here is the tricky part,  most saavy security consultants will apply 
normal security methodology techniques in examining a network using
<insert 
your favorite network topology mapping tool> and < insert your favorite 
network scanning tools> to assess the network.  In a previous life, I 
worked with a Phd who didn't want to listen that wrote some a
methodology 
for security assessments, only a minimum of what he wrote applies in 
examining a network with VoIP.

2. Is a network ready for VOIP?  That is an interesting question since
most 
<insert you favorite scanning tool here> will provide an organization or

security consultants very minimal information on whether a network is
ready 
for VOIP.  WARNING: If a security consultant who offers a VOIP readiness

check, inquire what tools they use, if their answer begins with <insert 
your favorite network scanning tool>, be very afraid.

3.  VOIP Attack suite - there are rudimentary scanning tools out there
for 
assessing VOIP products, but does not encompass all the components of a 
VOIP setup.  Here is the issue, running a scan across IP phones will
cause 
users of a particular organization get a little miffed, since most IP 
phones do not have denial of service protection built-in, so that is 
out.  Another issue is that most common intrusion detection systems have

not incorporated VOIP protocol decodes into their products yet, there
are a 
couple of pattern matching signatures out there for Sn0rt but very few,
so 
at most, when running VOIP attacks on a VOIP network, the majority of
noise 
will be from the users and very little information will be gathered
about 
the VOIP products except OS banner collection, and port flapping.

hope this helps

/m


At 08:32 AM 10/28/2004, Volker Tanger wrote:

> Greetings!
>
> On Wed, 27 Oct 2004 11:28:51 +0200 Frederic Charpentier
> <fcharpen () xmcopartners com> wrote:
> > does anyone have experiences or papers on VoIP pentest/assessment ?
> > Expecting classic OS/Network audits and H323/ASN.1 flaws, I can't
find
> > any documentations or papers about flaws in VoIP architecture.
>
> VoIP (SIP and H.323) do media transfer via (unencrypted) RTP/RTCP.
> SIP is a simple, unauthenticated cleartext protocol. H.323 similar
> (binary and more complex, but still unauthenticated).
>
> With ARPspoofing etc. it is simple to listen to voice streams or call
> setup - or change it. So re-routing voice streams or calls should be
> simple.
>
> Quite a high percentage of systems were/are susceptible to buffer
> overflows it seems (forgot the URL - about half a year ago).
>
> For other fun with SIP see e.g.
> http://www.infoanarchy.org/story/2004/9/15/23127/3363
>
> Bye
>
> Volker Tanger
> ITK Security