[VOIPSEC] Governments employing MiTM attacks against SSL

[J. Oquendo]

Ronald del Rosario wrote:

> > VoIPSA readers,
> >
> >  
> >
> > I would like to get your opinion regarding Christopher Soghoian and Sid
> > Stamm's recently release Proof of Concept (PoC) paper "Certified Lies:
> > Detecting and Defeating Government Interception Attacks Against SSL
> > <http://files.cloudprivacy.net/ssl-mitm.pdf> ".  
> >
> >  
> >
> >   
>   

This attack/technique and or concept isn't anything new. In fact, I can
point back to 1999 - 2000 when *someone* started a rumor about an agency
having this capability. Back then (for those who remember) it would have
been "gibberish" to flat out say: "Well... They're doing this!" because
the "security" world would have asked for definitive proof outside of a
*rumor* ...zzzzz Zorro(and other "rumors) were well-known - theories.
Bottom line... Deal with it. Seriously.

Even if you attempted to be vigilant about EVERY connection you could
make. Do you honestly and sincerely believe you will be opening
(manually) every single cert when you visit an HTTPS enabled site? How
long before you get tired. Even if you did, netsed is a PITA. Do you
honestly think you could avoid this. You could theoretically try...
Tunnel into one machine, into another into another... How long before
you get tired?

If you have zero to hide, you have zero to worry about. The logical
truth is, so much data need be sifted through, unless - again - you had
something to hide, the odds of the government using YOUR data from a
MITM attack is highly unlikely. In fact, you could probably hit the
Powerball in every state before your data made even a blip on the radar.


> > I understand CALEA, but when a government can easily fool an end-user by
> > employing a MiTM method and present a fake signed Digital Certificate
> > for its own purpose, how can we accelerate the adoption of Cloud
> > Computing and VoIP in the industry and build trust when more and more
> > news describing how flawed the underlying technology designed to protect
> > their transaction is being bypassed?
> >
> >   
>   

What does this have to do with cloud computing exactly? The worst thing
you could seriously ask for as a fix in any case, is to place your trust
on a cloud services provider. It goes right back to the core of this
message... MITM. Do you think for a second that under a National
Security Letter a cloud provider wouldn't do the same?

As for fixes... There was some work done to defend against MITM (SRP
(RFC2945), SPEKE, 1363 (http://grouper.ieee.org/groups/1363/)) but the
general public doesn't want secure (seriously). They want it fast...
Right now... "What the heck is this password?! Stupid machine... I'll
make the password: drowssap no one will ever guess!" or "BofA... Oh look
a picture of a guitar... I'll make my sitekey Gibson!... No one will
ever know!" No matter what you personally think, superlargemegaco's have
to cater to the majority. And the majority say: "Huh? What is this
certificate error message?... Oh well I don't know I guess I'll accept!"

No matter what you would like, the honest/realistic answer is... 1)
Cloud Security is as secure as you giving me your machine... Even if I
were say your best friend, doesn't mean I'm competent. Therefore when it
comes to security, its better to take care of your own security... You
have more to lose than a cloud provider... 2) There are fixes for MITM
... Don't hold your breathe on it. Plenty of certificate generating
monopolies with deep pockets for lobbying abound. 3) If you have nothing
to worry about... Why bother making noise?

Lastly... Unless you are your own carrier, own your own network (pipes
and all...) There is no way you can stop something like Naurus or Packet
Forensics... Besides, unless you actually took the time to READ PF's
data about the five series, you shouldn't worry: "users have the ability
to import a copy of any legitimate key they obtain legally - or they can
generate their own look-a-like keys" ... Translation: Either the gets a
hold of a valid cert (which in that case, no special device is really
needed anway) OR they create a fake cert and a user (as stated above)
says: "What the! Who! What is this certificate warning?! Better press
ok, I need to do my banking fast!"



-- =+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ J.
Oquendo SGFA, SGFE, C|EH, CNDA, CHFI, OSCP "It takes 20 years to build a
reputation and five minutes to ruin it. If you think about that, you'll
do things differently." - Warren Buffett 227C 5D35 7DCB 0893 95AA 4771
1DCE 1FD1 5CCD 6B5E
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x5CCD6B5E