Encrypted Magic Folders

["Alex Woolfson" <abdiel@worldnet.att.net>]

Hello!

First of all let me go on record by saying that I really like the idea and
the interface you've designed for your product, Encrypted Magic Folders.
The interface is easy to use and efficient.  Job well done!

There are, however, some security concerns I'd like resolved before I would
be interested in purchasing your product.

You scramble filenames which is an excellent idea.  I might suggest that you
also scramble the directory names as well.  When using the freeware program
Directory Snoop 1.1 designed by Kent Briggs (www.briggsoft.com) in Windows
95, I was easily able to find the directories that EMF had hidden.   I
wasn't able to access them, but I could see they were there *and also what
their names were*.  I don't know if it is possible to hide these directories
from such a program (although, obviously, that would be ideal), but it seems
that we should at least be able to scramble the directory filenames so that
little information is gleaned about which directories are worth attacking by
cryptanalysis and which are not worth the attackers' trouble.

On the question of encryption algorithms, let me add my voice to those who
have been suggesting that you use strong and tested algorithms like IDEA or
Blowfish.  I feel very strongly that this is essential to any product that
purports to provide true security to its users.  If your product were meant
to merely keep your kid sister from reading your files, then the security
through obscurity your product seems to offer would be sufficient.  But your
marketing materials suggest that you wish to offer a greater level of
security than that.

Not being a cryptologist, I consulted with members of the Cypherpunk mailing
list to aid me in evaluating the security of your software referring them to
the "How Secure Is It?" section of your Help file.  These guys really know
their stuff.  The verdict was unanimous--that the security of Encrypted
Magic Folders was not sufficient for anything but the most casual uses.  I
have included below one of the more thorough and detailed of the replies in
the hopes that it will be useful in improving the strength of your software.

I would also feel more comfortable if you were willing to have the source
code of your product reviewed by industry experts (such as Bruce Schneier of
Counterpane Systems who wrote the very well respected Applied Cryptography
book as well as Blowfish).  If you are concerned about someone stealing your
idea, non-disclosure agreements should be able to offer sufficient legal
protection.  Truly secure algorithms and programs can withstand this kind of
scrutiny without compromising their security.  Relying on the fact that your
company is "too small and obscure" for an attacker to develop an prepackaged
attack offers little comfort--if someone really wants to have access to my
files, I imagine they'll devote the time to attempt to crack your
product--copyright infringement notwithstanding!

If you haven't already, I would strongly suggest that your developers read
Schneier's book Applied Cryptography.  It does a much better job of
explaining the difference between true security and security through
obscurity.  And it's written for programmers who wish to use cryptography in
their programs.  It's not a bad read, either.  :-)

I am really interested in purchasing a product like Encrypted Magic Folders,
but for such an application, security is obviously the biggest priority.  I
believe others who have not taken the trouble to write you feel the same
way.  If EMF makes use of strong encryption and is willing to have the
source code of its product reviewed by industry experts, I would be
interested in paying for it.  For a $60.00 product, I believe it is fair to
expect this kind of protection.

Yours

Alex Woolfson

Below is the response of Ray Arachelian of the Cypherpunks list.  There is a
bit of sarcasm involved, but I trust you'll be able to see beyond it to the
valid criticism he makes.  I hope you find it useful.

On Thu, 4 Dec 1997, Alex Woolfson wrote:

>     EMF's encryption offers good protection and excellent speed.  It
>     hasn't been broken yet.  It is, as far as we know, exportable.  THERE
>     IS NO BACKDOOR.  Should you forget your password there is nothing we
>     can do to decrypt your encrypted files.

If it is exportable, then it is weak encryption by definition.  The
question is how weak.  Reading their marketting crap provides more
insights in just how lame this is:

>     Quite a few people ask us how big EMF's key size is.  They've learned
>     from other encryption programs that the bigger the key the stronger
>     the encryption.  This really doesn't apply to EMF.
>
>     We developed our own encryption instead of using a standard because
>     we wanted EMF to be able to decrypt at the byte level.  In this way
>     we only need to decrypt/encrypt the data your programs require and
>     not the entire file.

In other words: security through obscurity.  Smells like snake oil, looks
like snake oil, it even has bits of snake scales in it.  Heck, if they say
that the key size is not important, it's likely very tiny, or regardless
of what key size you'd use (if you had any choice), it's such a weak
cypher that it wouldn't get you anything.


>     In theory, because we decrypt at the byte level, the biggest key we
>     could use would be 8 bits - which is a joke.  So instead of
>     decrypting every hunk of data using the same key, as most other
>     encryption programs do, we developed an algorithm to vary the key
>     based on the data's location within the file.  In this way we get
>     both high security and high speed.  We are trying to patent EMF's
>     encryption method.

Whoop.  so the key is the location of the data?  At most, this means a 32
bit key on most file systems, 64 bits on newer file systems.  This assumes
that your file is long enough for that.  But as a key, this is totally
useless as the key is the location which is visible to all.  Likely they
would be wise to "mix" this location key with the passphrase.

This offers little more than CBC would.  The only difference is that CBC
is based on the previous block, where this is based on the location of a
byte in the file (which is known) whereas a CBC depedns on knowledge of
the previous block.

If they did it the right way, they'd use the passphrase to build a lookup
table of large keys based on the byte location within a file, but because
this depends on your passphrase, your security is still that of the
passphrase at maximum.  There may be ways to weaken this.

>     Having said all that, truth is, most encryption isn't "cracked" by
>     breaking the algorithm, it's done by guessing the password.  Brute
>     guessing of passwords tends to level the playing field tremendously.
>     We actually have an advantage because we aren't an established
>     standard.  Because we're small and relatively obscure chances are no
>     one will take the effort to write a password guessing program (which
>     incidentally would violate copyright and intellectual property laws.)
>     Even if someone were to go thru all this effort we could easily
>     change the encryption method for the next update.

This is total bullshit.  It's Grade A, government approved exportable
snake oil.  Just because they are small and obscure it doesn't guarantee
you any extra safety.

Just because nobody has bothered to write a brute forcer for their code
doens't mean nobody can do so.  And no, this would not violate any of
their copyright if the author of the brute forcer proggie wrote it without
copying their code.  It might violate intellectual property laws if they
had a patent on it, but they can't get one since they're going by
security through obscurity.  They claim to be applying for a patent, but
if they do, their code is published and their weaknesses will be
displayed.

It might violate trade secret laws if someone who has the source or the
algorithm shares it, but not if someone reverse engineers it (depending on
reverse engineering laws of course.)

Just because they could "rewrite" the code in the next update against
cracking or brute force attacks does not buy you security.  FYI: If you
use this crap and someone steals a copy of the encrypted files off your
hard drive, it doesn't matter what new updates this company produces.  The
fact is you used the older breakable version, the attacker has your
cyphertext and will decode it.


>     If we used an established encryption method like DES or Blowfish then
>     your files would probably have to be fully decrypted when opened,
>     would exist on disk as unencrypted while you're using them, and then
>     would need to be encrypted when closed.

Translation: We don't use CBC, because that would be more work and slow
down our crappy cypher.  Deeper translation: break any portion of any
encrypted file and you'll likely break the whole thing if not huge parts
of it.

"Established encryption methods (sic)" use CBC because it adds security.
Their algorithms are well published and well known so people can find
their weaknesses and publicise them.  Whatever weaknesses there are in
this thing are hidden from those who can't or won't reverse engineer it.
But there are plenty of those who can and will, and have no qualms about
releasing it anonymously, or worse: keeping the expolit to themselves so
as to exploit those stupid enough to use this lame shit.

It only takes one cracker with a good disassembler to reverse engineer
their code and find all the holes, and they'll be out of business.  If
what protects your data is the cypher, and not the key, then breaking the
cypher is all you need to do.

Heck, I'd bet they use something really shitty like this: (since "key size
is not important", and "depends on location".)


char code(char *passphrase, char data, long location);
{
  int i; char c;

  for (i=0; passphrase[i]>0; c^=passphrase[i++]);

  return (data ^ c ^ (char)(location & 0xff));
}

Gee, I probably violated their "intellectual" property by guessing their
code, heh.


> This has multiple
>     disadvantages.  First, if your computer shuts down while you have
>     "encrypted" files open, then those files would be unencrypted.  This
>     doesn't happen with EMF as your encrypted files are always encrypted
>     as stored on disk.  The second disadvantage is that it slows things
>     down tremendously.

Not if you use an encrypted disk driver.  then, all your data is encrypted
all the time.  (You get into other issues such as keeping the passphrase
or key from falling into the pagefiles, etc...)

[Meaningless babble deleted.]



=====================================Kaos=Keraunos=Kybernetos==============
.+.^.+.|  Ray Arachelian    |Prying open my 3rd eye.  So good to see |./|\.
..\|/..|sunder@sundernet.com|you once again. I thought you were      |/\|/\
<--*-->| ------------------ |hiding, and you thought that I had run  |\/|\/
../|\..| "A toast to Odin,  |away chasing the tail of dogma. I opened|.\|/.
.+.v.+.|God of screwdrivers"|my eye and there we were....            |.....
======================= http://www.sundernet.com ==========================