Re: Guidance Software response to iSEC report on EnCase (fwd)
> they were able to identify six test scenarios, out of =93tens of thousand=
s=94 of test scenarios run
It only takes one. Plus this statement implies that its relatively hard to
crash EnCase; anyone thats used it on any regular basis knows that
crashing EnCase is on par with kicking retarded children out of their
wheelchairs. I will admit that I don't have any examples off-hand though-
Cory Altheide, please stand up.
> , that apparently revealed
> minor bugs
By minor, you mean things like (1) where a disk image cannot be acquired
or (2) that appears to cause an out-of-bounds memory operation or (3)
which most likely has one hell of a race condition?
> All of the testing involved
> intentionally corrupted target data that highlighted a few relatively
> minor bugs.
Yes, and pretty much every exploit on the planet involves intentionally
corrupted data.
> The issues raised do not identify errors affecting the
> integrity of the evidence collection or authentication process, or the
> EnCase Enterprise process (i.e., the operation of the servlet code or
> the operation of the SAFE server). Moreover, the issues raised have noth=
ing to do with the security of the product.
Yes, the md5 hash of the lack of an image is 0, so no data was corrupted
(or acquired). Furthermore, are you will to step out and say that none of
these bugs will get an examiners workstation owned?
But hey, inability to acquire an image and dangling pointers are not a
security condition.
> Therefore, we strongly dispute any media reports or commentary that
> imply that there are any =93vulnerabilities=94 or =93denials of service=
=94 exposed by
> this report.
Of course you do, I can't blame you or your company. But let's be serious
here for a moment, wishing that you're the queen of England doesn't make
it so.
> Forensic examiners will inevitably come across corrupted data on target s=
ystems from time to time; and in standard computer forensics training, incl=
uding classes offered by Guidance Software, examiners are trained to accoun=
t for such issues. In addition, while Guidance Software maintains a robust =
in-house quality assurance process and strives to make our software as stab=
le as possible, no software is completely crash-proof and there will always=
be anomalies, particularly involving extreme scenarios of corrupted target=
data.
Did you really just turn the shoddiness of your application into a
training opportunity?
> 1.=09[Logical] Disk Image Cannot be Acquired With Certain Corrupted MBR P=
artition Table.
>
> Response: It should be no surprise to any computer forensic examiner that=
a logical copy of a volume may not be possible if that volume has a corrup=
ted MBR Partition table. EnCase features an option to acquire the target me=
dia physically, rather than logically, to specifically account for this typ=
e of scenario. The authors ignored the option of acquiring the data physic=
ally. Also, by corrupting the MBR Partition table, the perpetrator would l=
ikely render his computer inoperable, which calls into question both the li=
kelihood and feasibility of such a tactic.
So if I give you a computer with a corrupted MBR partition table that
still boots, what are you going to give me? You realize that the corrupted
MBR partition table really doesn't break anything, it just requires that th=
e boot code in the first
sector of the disk to be rewritten. Do you actually understand how a PC
boots?
Here's a good idea. corrupt the partition table, store the real one at
some distant offset, rewrite bytes 0-446 in the first sector to load the
sector with the 'real' partition table, read it and boot correctly.
Furthermore, I know some BIOSs will still boot without a valid MBR
partition table in the first place.
> 2. Corrupted NTFS file system crashed EnCase during acquisition.
>
> Response: The authors state that =93this issue appears to be caused by an=
attempt to read past the end of the buffer.=94 However, EnCase features a=
n option to de-select the automatic reading of the file system during the a=
cquisition process. Thus, there is an easy work-around. Also, by corruptin=
g the NTFS partitions, the perpetrator would likely render his file system =
dysfunctional, which calls into question both the likelihood and feasibilit=
y of such a tactic. Thus, the chances of this specific scenario occurring =
in the field are extremely remote; however, Guidance Software will test and=
, if verified, place this anomaly in its development queue to address the c=
rashing problem in the future.
So really all I need to do is wrap my partition/file-system in a
corrupted NTFS (btw NTFS file system is redundant), and poof I potentially
owned the forensics workstation and with a little lucky just kept you from
accessing my data.
Furthermore, okay, seeing as I control the disk, I potentially control
whats on both sides of the buffer you read past, so I can potentially
corrupt the integrity of the image you made, and thats even I can't grab
control of eip through you're bad read.
> 3.=09Corrupted Microsoft Exchange database crashes EnCase during multi-th=
readed search/analysis concurrent to acquisition
>
> Response: The report discloses that this particular anomaly occurred only=
when every single check box was selected in the search dialogue box, inclu=
ding the search, hash value calculation and verify file signatures features=
=2E This means that EnCase was directed to acquire an Exchange database and=
perform a detailed multi-threaded search and analysis of the data at the s=
ame time. This procedure is extremely inconsistent with best practices and =
akin to opening several hundred files in a word processing program, which o=
f course would cause a memory overload.
So, you have options that you don't expect customers to select? If this is
such a problem, why do you allow all of the options to be selected at the
same time?
> 5. EnCase Had Difficulty Reading Intentionally Corrupted NTFS File System=
Directory.
>
> Response: This issue involves the authors intentionally corrupting an
> NTFS file system to create a =93loop=94 by, =93replacing a directory entr=
y for
> a file with a reference to the directory=92s parent directory.=94
> Experienced forensic examiners are trained to identify such instances of
> data cloaking. The purposeful hiding of data by the subject of an
> investigation is in itself important evidence and there are many
> scenarios where intentional data cloaking provides incriminating
> evidence, even if the perpetrator is successful in cloaking the data
> itself. The chances of this specific scenario occurring in the field are
> extremely remote, but Guidance Software will test and, if verified,
> place this anomaly in its development queue to be addressed in the future=
=2E
Your argument is that they should be able to detect this
manually? Isn't that why they bought your overpriced software in the first
place?
why are the odds of this remote? Because you said so, or because of
reasons like the corrupted MBR partition table?
So here's some truth, you're a forensics software company, one that deals
by and large with LEO, which means you have an incredibly high bar to
fill, if you're application is not pristine, all sorts of 'bad people with
good lawyers (tm)' walk the streets, essentially invalidating the entire
purpose of your tool. Stating that there are work-arounds or attempting to
sweep bugs under the carpet just exasperates the situation.
討論串 (同標題文章)
完整討論串 (本文為第 1 之 2 篇)：