Forensic Examination of a Floppy Disk
Analysis of The Honeynet Project's Scan 24

Marc Bayerkohler
Vince Elliott

October 25, 2002

Introduction

The Honeynet Project releases monthly challenges in which raw data from a security incident is provided for analysis.  Scan 24, the challenge for October 2002, is a fictional situation based on the arrest of a suspected drug dealer.  A single floppy disk contains evidence that the police need, but the suspect has tried to hide.  

To complete the challenge, a forensic examination of the disk must be performed, and the challenge questions answered.

File Recovery

Initial data gathering

The image of the floppy disk was downloaded onto a Red Hat 7.2 linux box, and the md5sum compared to that from the website.
$ wget http://old.honeynet.org/scans/scan24/image.zip
--17:01:02--  http://old.honeynet.org/scans/scan24/image.zip
           => `image.zip'
Connecting to www.honeynet.org:80... connected!
HTTP request sent, awaiting response... 200 OK
Length: 18,146 [application/zip]

    0K .......... .......                                    100% @   5.26 KB/s

17:01:05 (5.26 KB/s) - `image.zip' saved [18146/18146]

$ md5sum < image.zip
b676147f63923e1f428131d59b1d6a72  -
The Honeynet Project's web site confirms the MD5sum:
image.zip MD5 = b676147f63923e1f428131d59b1d6a72 ( image.zip )

The first command run was the unix 'strings' command, which displays any text strings that it finds in a file.  
$ strings image | more
MSDOS5.0
)ÏͱÄNO NAME    FAT12   3É
Ѽð{
|8N$}$
<snip>
This revealed a wealth of information.  The disk identifies itself as "MSDOS5.0" and is formatted as FAT12, apparently on a WindowsNT system, because of the reference to the "NTLDR".  

File names were found in the directory table (see Table 1), including one deleted file.  The text of a letter was recovered (Listing 1), what appears to be a password (Table 2), and reference to a spreadsheet (Table 2).

åIMMYJ~1DOC
A deleted file, possible a Microsoft Word document
COVERP~1JPG
A possible JPEG image file
SCHEDU~1EXE
A possible executable file
Table 1: File names in the directory table.

Jimmy Jungle
626 Jungle Ave Apt 2
Jungle, NY 11111

Jimmy:
Dude, your pot must be the best it made the cover of High Times Magazine! Thanks for sending me the Cover Page. What do you put in your soil when you plant the marijuana seeds? At least I know your growing it and not some guy in Columbia.

These kids, they tell me marijuana isnt addictive, but they dont stop buying from me. Man, Im sure glad you told me about targeting the high school students. You must have some experience. Its like a guaranteed paycheck. Their parents give them money for lunch and they spend it on my stuff. Im an entrepreneur. Am I only one you sell to? Maybe I can become distributor of the year!

I emailed you the schedule that I am using. I think it helps me cover myself and not be predictive.  Tell me what you think. To open it, use the same password that you sent me before with that file. Talk to you later.
Thanks,
Joe
Listing 1, Text of a letter.

pw=goodtimes
A possible password
Scheduled Visits.xls
Name of a Microsoft Excel spreadsheet
Scheduled Visits.xlsPK
Name of a Microsoft Excel spreadsheet
Table 2, Text fragments in the image file.

To access the files in the image of the floppy, it was linked to a loop device and mounted on the linux filesystem.  The loop device was mounted with a number of options (Table 3) to ensure the integrity of the data, and protect the server.

# /sbin/losetup /dev/loop0 image
# mount -o ro,nodev,nosuid,noexec /dev/loop0 mount/
# mount
...
/dev/loop0 on /tmp/sotm/24/mount type vfat (ro,noexec,nosuid,nodev)

And it could now be accessed from the linux filesystem.
$ cd mount
$ ls -l

total 17
-rw-r--r--    1 root     root        15585 Sep 11 08:30 cover page.jpgc
-rw-r--r--    1 root     root         1000 May 24 08:20 schedu~1.exe

ro
Read only
nodev
Do not interpret character or block special devices  on  the file system
nosuid
Do not  allow  set-user-identifier  or  set-group-identifier bits to take effect
noexec
Do not allow execution of any binaries on the mounted file system
Table 3, Mount options for the loop device.

The unix 'file' command attempts to classify a file by looking at its data.  For instance, a JPEG file could be renamed to 'command.com', but it would still have the JPEG headers in the file to identify it.

$ file *
cover page.jpgc           : PC formatted floppy with no filesystem
schedu~1.exe:               Zip archive data, at least v2.0 to extract
The "cover page.jpgc" doesn't say it is a JPEG file, but the text of the letter seems to indicate it should be a valid image file, so perhaps it has been altered (encrypted/encoded/wiped) to conceal the picture.  The file cannot be displayed with image viewing programs.  According to 'file', "schedu~1.exe" appears to be a self-extracting archive in the Zip format.  Since the longer filename "Scheduled Visits.xls" was also found, this is probably the full name of the archived file.  However, if run as an executable, it errors out, and zip archive utilities (WinZip 8.0, unix unzip) are not able to make sense of it.  This file may also have been changed to conceal its data.  "åIMMYJ~1DOC" has been deleted.  These three files must be recovered to a readable format for further investigation.

Ashes to ashes, disk to disk

To continue the examination, it was useful to restore the image to a floppy disk.  This was done from the command line on a Windows box using 'fdimage.exe', available as part of OpenBSD.  
C:\scan24>fdimage image a:
The disk was physically write-protected and scanned for viruses (negative).  Writing was then allowed, and reimaged as needed during the investigation.

Recovering Jimmy Jungle.doc

On a FAT file system, when a file is deleted, only the directory listing for the file is changed.  The first character of the filename is changed, its location is marked as unused, and the file is no longer visible.  The data of the file itself is left on disk, and would eventually be over written in the normal course of moving data on the disk.  There are many utilities designed to recover files deleted in this manner.  "åIMMYJ~1DOC" appears to be a deleted Word document.

Windows95 includes the 'undelete' command, and was used to recover the file. after booting to MS-DOS mode.
C:\>undelete a:*.*

MS-DOS directory contains 1 deleted files.
Of those, 1 files may be recovered.

Using the MS-DOS directory method.

?IMMYJ~1 DOC 20480  4-15-02  2:42p  ..A  Undelete (Y/n)?y
Please type the first character for ?IMMYJ~1.DOC: j

File successfully undeleted.
This fully recovered the Jimmy Jungle.doc file.

Recovering Scheduled Visits.xls

As previously mentioned, although 'file' indicates "SCHEDU~1.EXE" is a zip file, zip utilities cannot read the archive.
$ unzip schedu~1.exe
Archive:  schedu~1.exe
  End-of-central-directory signature not found.  Either this file is not
  a zipfile, or it constitutes one disk of a multi-part archive.  In the
  latter case the central directory and zipfile comment will be found on
  the last disk(s) of this archive.
note:  schedu~1.exe may be a plain executable, not an archive
unzip:  cannot find zipfile directory in one of schedu~1.exe or
        schedu~1.exe.zip, and cannot find schedu~1.exe.ZIP, period.
In the strings output "Scheduled Visits.xlsPK", the "PK" is another indication that this is a zip file.  According the the zip file format found at My File Formats, PK, or 50 4B in hex, is used to seperate different sections in a zip file.  WinHex is a full featured windows based hex editor by X-Ways Software Technology AG, which provides a free download for evaluation.  On a Windows2000 computer, "SCHEDU~1.EXE" was opened using WinHex.  According to the zip format, PK should appear three times, but it only occurs once in this file, which ends suddenly without the proper signatures for the end of a zip archive.  The disk image was opened in WinHex, and a search for PK shows all three expected hits, including a proper zip file ending.

Following the zip archive format, the file should start at offset 0x0000D000:
Offset      0  1  2  3  4  5  6  7   8  9  A  B  C  D  E  F

0000D000   50 4B 03 04 14 00 01 00  08 00 98 5A B7 2C C7 55   PK........˜Z·,ÇU
0000D010   60 8D EA 08 00 00 00 42  00 00 14 00 00 00 53 63   `�ê....B......Sc
0000D020   68 65 64 75 6C 65 64 20  56 69 73 69 74 73 2E 78   heduled Visits.x
0000D030   6C 73 94 C8 31 2A E3 49  0B DB A8 10 C2 70 9D FC   ls”È1*ãI.Û¨.Âp�ü
and end at offset 0x0000D973:
0000D940   00 00 20 00 B6 81 00 00  00 00 53 63 68 65 64 75   .. .¶�....Schedu
0000D950   6C 65 64 20 56 69 73 69  74 73 2E 78 6C 73 50 4B   led Visits.xlsPK
0000D960   05 06 00 00 00 00 01 00  01 00 42 00 00 00 1C 09   ..........B.....
0000D970   00 00 00 00                                        ....
Using WinHex, this was tested.  We defined a block with the correct offsets, and saved it to a new file.
Edit -> Define Block...
<image>
Edit -> Copy Block -> Into New File
The new file was a valid zip archive, and could be opened.  The password needed to open the file is the one found earlier (Table 2) with 'strings', and mentioned in the letter (Listing 1).

This proves that the entire archive is on the disk, but that accessing it normally does not give the entire file.  A directory listing shows the file size is suspicious.
05/24/2002  08:20a               1,000 SCHEDU~1.EXE
The file size is exactly 1,000, an unlikely figure.  The recovered file is 2420 bytes in length.  This means that the file length listed in the File Allocation Table (FAT) is incorrect.  Given that the length was changed to an even number, it may have been edited manually using a hex editor to make the file unreadable until the length is changed back.

DOS Drive Tool (DDT) by Dan Heimsoth is a directory/sector editor available for free download.  To test the method the suspect would use to access the file, we changed the file size from 1000 bytes (0x3E8) to what we now know is the correct length, 2420 bytes (0x974).  After writing the change to disk (F9), the entire file was usable from the floppy, giving complete access to "Scheduled Visits.xls".

  PATH = A:\
  ENTRY # = 0008
┌───────────────┬────────────────────────────────────────────┬─────────────────┐
│  "|"= unused  │            Vol     Rd.only  [ Attributes ] │ Start Cluster   │
│  "+"= erased  │         Sub │    Hid │ ┌─────────┬─────────┤     ┌───────────┤
│  Filename.ext │      Arc │  │ Sys │  │ │   Date  │   Time  │     │ File Size │
├───────────────┴────────────────────────┴─────────┴─────────┴─────┴───────────┤
│  +d|o|c||.|?? │ 0  0  0  0  1  1  1  1 │ 15-31-07│ 31:63:62│ 0000│ FFFFFFFF  │
│  +J|i|m|m.|y| │ 0  0  0  0  1  1  1  1 │  3-12-80│  0:03:14│ 0000│ 002E0065  │
│  +IMMYJ?1.DOC │ 0  0  1  0  0  0  0  0 │  4-15-02│ 14:42:30│ 0002│ 00005000  │
│  Bg|c| | .| | │ 0  0  0  0  1  1  1  1 │  1-00-80│  0:01:00│ 0000│ 00200020  │
│  ?c|o|v|e.|r| │ 0  0  0  0  1  1  1  1 │  1-14-80│  0:03:10│ 0000│ 0070006A  │
│  COVERP?1.JPG │ 0  0  1  0  0  0  0  0 │  9-11-02│  8:30:52│ 01A4│ 00003CE1  │
│  Bi|t|s|..|e| │ 0  0  0  0  1  1  1  1 │  1-00-80│  0:01:00│ 0000│ 00200020  │
│  ?S|c|h|e.|d| │ 0  0  0  0  1  1  1  1 │  2-22-80│  0:01:00│ 0000│ 00730069  │
│> SCHEDU?1.EXE │ 0  0  1  0  0  0  0  0 │  5-24-02│  8:20:32│ 0049│ 000003E8 <│
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
└──────────────────────────────────────────────────────────────────────────────┘
[  F1  ][  F2  ][  F3  ][  F4  ][  F5  ][  F6  ][  F7  ][  F8  ][  F9  ][ F10  ]
  HELP   SEARCH   EXIT     FAT   CLUSTR  SUBDIR                   WRITE   CLEAR
F1 = Display Help information

Table 4, Using DOS Drive Tool to edit the FAT.

A directory listing now shows the correct file length.
05/24/2002  08:20a               2,420 SCHEDU~1.EXE

Recovering cover page.jpg

A directory listing of the floppy shows a "cover page.jpgc".
09/11/2002  08:30a              15,585 cover page.jpgc
But the file is difficult to access.  Most windows programs are unable to access it, and the command line tools do not do any better.
C:\dloads\sotm24\temp>copy a:c*.* .
a:cover page.jpgc
The system cannot find the file specified.
        0 file(s) copied.
On the linux system, a directory listing shows that the filename has been padded to make it more difficult to access.
$ ls --escape
cover\ page.jpgc\ \ \ \ \ \ \ \ \ \ \
The 'hexdump' command was used to view the contents of cover page.  Inside is nothing but 0xF6 and 0x00.
$ hexdump cover\ page.jpgc\ \ \ \ \ \ \ \ \ \ \
0000000 f6f6 f6f6 f6f6 f6f6 f6f6 f6f6 f6f6 f6f6
*
0000200 0000 0000 0000 0000 0000 0000 0000 0000
*
0003ce0
This is certainly not a file that follows the JPEG file format.  According to the format, a JPEG file should start with 0xFFD8, and end with 0xFFD9.  Searching the image with WinHex reveals both of these exist, and define a block that looks like a valid JPEG file.  Using the same procedure as above, WinHex was used to save the block to a new file, which was indeed a JPEG image file.

The JPEG file was on the floppy, but the directory was misdirected to point at another part of the disk.  The suspect must want the file, because pains have been taken to conceal it, rather than delete it.  Therefore it must be possible to reverse the process to access the file normally.

Using WinHex to open the floppy disk (not the image) shows that the actual JPEG file starts at cluster 42, which the directory lists as free space.  

Screenshot of WinHex

DDT shows "COVERP?1.JPG" starts at cluster 0x01A4 (see Table X), or 420 decimal.  Likewise, WinHex shows that cluster 420 belongs to "cover page.jpgc           ".      

Screenshot of WinHex

420 is an easy change from 42; it appears the suspect edited the FAT again, this time changing the starting cluster of a file.  DDT was used to correct the starting cluster of "cover page.jpgc" to 42 (0x2a).

  PATH = A:\
  ENTRY # = 0005
┌───────────────┬────────────────────────────────────────────┬─────────────────┐
│  "|"= unused  │            Vol     Rd.only  [ Attributes ] │ Start Cluster   │
│  "+"= erased  │         Sub │    Hid │ ┌─────────┬─────────┤     ┌───────────┤
│  Filename.ext │      Arc │  │ Sys │  │ │   Date  │   Time  │     │ File Size │
├───────────────┴────────────────────────┴─────────┴─────────┴─────┴───────────┤
│  +d|o|c||.|?? │ 0  0  0  0  1  1  1  1 │ 15-31-07│ 31:63:62│ 0000│ FFFFFFFF  │
│  +J|i|m|m.|y| │ 0  0  0  0  1  1  1  1 │  3-12-80│  0:03:14│ 0000│ 002E0065  │
│  +IMMYJ?1.DOC │ 0  0  1  0  0  0  0  0 │  4-15-02│ 14:42:30│ 0002│ 00005000  │
│  Bg|c| | .| | │ 0  0  0  0  1  1  1  1 │  1-00-80│  0:01:00│ 0000│ 00200020  │
│  ?c|o|v|e.|r| │ 0  0  0  0  1  1  1  1 │  1-14-80│  0:03:10│ 0000│ 0070006A  │
│> COVERP?1.JPG │ 0  0  1  0  0  0  0  0 │  9-11-02│  8:30:52│ 002A│ 00003CE1 <│
│  Bi|t|s|..|e| │ 0  0  0  0  1  1  1  1 │  1-00-80│  0:01:00│ 0000│ 00200020  │
│  ?S|c|h|e.|d| │ 0  0  0  0  1  1  1  1 │  2-22-80│  0:01:00│ 0000│ 00730069  │
│  SCHEDU?1.EXE │ 0  0  1  0  0  0  0  0 │  5-24-02│  8:20:32│ 0049│ 00000974  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
│  ||||||||.||| │ 0  0  0  0  0  0  0  0 │  0-00-80│  0:00:00│ 0000│ 00000000  │
└──────────────────────────────────────────────────────────────────────────────┘
[  F1  ][  F2  ][  F3  ][  F4  ][  F5  ][  F6  ][  F7  ][  F8  ][  F9  ][ F10  ]
  HELP   SEARCH   EXIT     FAT   CLUSTR  SUBDIR                   WRITE   CLEAR
F9 = Write the new DIRECTORY data to the disk.

Table 5, DDT showing the FAT after corrections.

Mounting the floppy under linux makes it easy to evade the strange file name.
$ cp cover\ page.jpgc\ \ \ \ \ \ \ \ \ \ \  /tmp/coverpage.jpg
And now we have recovered the "cover page.jpg".  It should be noted that Symantec's Norton Disk Doctor was also able to recover this file, even though the directory entry was completely wrong, and the actual file was in space listed as free.

Below is a simplified graphic showing the differences between the actual file locations and those described by the FAT edited to conceal them.

Graphic of file placement on disk
Figure 1, Placement of files on the disk.

Recovered files

The files recovered can be accessed here.  The MD5 is provided to assure the integrity of the evidence to be used in court.

Filename
MD5
Size
Description
JIMMYJ~1.DOC
b775eb6a4ccc319759d9aaae1e340acc
20480
Letter from Mr.Jacobs to Mr. Jungle.  MS Word document.
SCHEDU~1.EXE
ab6a87ff7acf36eb2803dfd12ec2036f
2420
Zip archive of Scheduled Visits.xls.
Scheduled Visits.xls
63201eae843132476343a862966f7643
16896
Rotating schedule of visits to high schools.  MS Excel.
cover page.jpg
e30e8ecec4500678f7270e96b1d5663b
15585
Cover of a "High Times Magazine" declaring Mr. Jungle a featured pot grower.  JPEG image.
Table 6, Files recovered from the floppy disk.

File Analysis

Time stamps

Utilizing ils from The Coroner's Toolkit, inode information was retrieved from the image:
# ils -e -f fat12 /tmp/sotm/24/image.dd
class|host|device|start_time
ils|wghou899061a7|/tmp/sotm/24/image.dd|1035497126
st_ino|st_alloc|st_uid|st_gid|st_mtime|st_atime|st_ctime|st_mode|st_nlink|st_size|st_block0|st_block1
2|a|0|0|0|0|0|40000|1|7168|1|0
5|f|0|0|1018899750|1031720400|1031752188|100777|0|20480|2|0
8|a|0|0|1031751052|1031720400|1031752226|100777|1|15585|420|0
11|a|0|0|1022246432|1031720400|1031752238|100777|1|1000|73|0
Using the retrieved inodes and istat from the coroner's toolkit, some basic file information was returned:
# istat -f fat12 /tmp/sotm/24/image.dd 2
Directory Entry: 2
Allocated
DOS Mode: Directory, Read Only, Hidden
size: 7168
num of links: 1
Directory Entry Times:
Written:        Wed Dec 31 18:00:00 1969
Accessed:       Wed Dec 31 18:00:00 1969
Created:        Wed Dec 31 18:00:00 1969
Sectors:19 20 21 22 23 24 25 26 27 28 29 30 31 32
This appears to be the directory (or File Allocation Table) of the floppy.
# istat -f fat12 /tmp/sotm/24/image.dd 5
Directory Entry: 5
Not Allocated
DOS Mode: File
size: 20480
num of links: 0
Name: _IMMYJ~1.DOC
Directory Entry Times:
Written:        Mon Apr 15 14:42:30 2002
Accessed:       Wed Sep 11 00:00:00 2002
Created:        Wed Sep 11 08:49:48 2002
Sectors:33
This appears to be a deleted file! See the 'File Recovery' and 'Recovering Jimmy Jungle.doc' for details on the recovery. The 'Written' date corresponds to the time frame from the recovered schedule spreadsheet. Further, from the police report "Jacobs has been seen on numerous occasions hanging out at various local high school parking lots around 2:30pm" it is possible that the computer on which Jacobs created this file is in close proximity to Richter High School, the school Jacobs was to visit that day (according to the recovered schedule).
# istat -f fat12 /tmp/sotm/24/image.dd 8
Directory Entry: 8
Allocated
DOS Mode: File
size: 15585
num of links: 1
Name: COVERP~1.JPG
Directory Entry Times:
Written:        Wed Sep 11 08:30:52 2002
Accessed:       Wed Sep 11 00:00:00 2002
Created:        Wed Sep 11 08:50:26 2002
Sectors:451
This appears to be an image file (from it's filename). Using a hex editor confirms this (A JFIF header is visible in the image).
# istat -f fat12 /tmp/sotm/24/image.dd 11
Directory Entry: 11
Allocated
DOS Mode: File
size: 1000
num of links: 1
Name: SCHEDU~1.EXE
Directory Entry Times:
Written:        Fri May 24 08:20:32 2002
Accessed:       Wed Sep 11 00:00:00 2002
Created:        Wed Sep 11 08:50:38 2002
Sectors:104 105
From the filename, this appears to be a Windows executable. Using a hex editor, it appears to be a zip file. Perhaps it's a self-extracting archive.

Hiding text within image files

The password, 'goodtimes', was recovered using the unix 'strings' command.  However, because the suspect manipulated the FAT, it is difficult to know if the text is an artifact of a previous deleted file, or if it was in fact part of another file.  Its location on the disk is at the end of the JPEG file.  Some image file formats, such as JPEG and GIF, allow text to be added to the end of a file while remaining a valid image.

For example, the images below have text added to them.
$ cat anne.jpg hidden.txt > anne_text.jpg




$strings anne.jpg | tail -3
Z#_a
zx{w
ln6=


$strings anne_text.jpg | tail -3
zx{w
ln6=
Look GIR, a secret message!


$strings anne_text.gif | tail -3
,|dG
 ]p|U@
;Look GIR, a secret message!



Misdirection

It also appears that Mr. Jacobs was using misdirection in his letter to Mr. Jungle.  In the letter, he refers to Mr. Jungles appearance on the cover of "High Times Magazine".  However, the recovered image is the cover of "POT SMOKERS MONTHLY".  This may have been an attempt to divert attention to the grower on the cover of High Times Magazine if his letter were intercepted.


HighTimesMagazine
Recovered image file, cover of POT SMOKERS MONTHLY, date unknown.
Archived image file, cover of High Times Magazine for September, 2002.

Answers to Questions

Who is Joe Jacob's supplier of marijuana and what is the address listed for the supplier?

Jimmy Jungle
626 Jungle Ave Apt 2
Jungle, NY 11111

What crucial data is available within the coverpage.jpg file and why is this data crucial?

The image itself declares Jimmy Jungle to be "This month's featured pot grower, smoker and seller".  This corroborates both Mr. Jacobs initial statement to police, which he had refused to validate, "Look at this stuff, Colombians couldn't grow it better! My supplier not only sells it direct to me, he grows it himself.", and his statements to the same effect in his letter to Mr. Jungle.  This could be crucial in preparing a case against Mr. Jungle even if Mr. Jacobs refuses to testify.

The suspect had manually edited the FAT, and so it is unclear exactly where the coverpage.jpg file ends on the disk.  It is possible that the password, 'goodtimes', was being stored as text on the end of the file.  This was crucial in unzipping the spreadsheet, although there are tools available for cracking encrypted zip archives.

What (if any) other high schools besides Smith Hill does Joe Jacobs frequent?

As listed in his Scheduled Visits spreadsheet: Birard High School, Hull High School, Key High School, Leetch High School, and Richter High School.

For each file, what processes were taken by the suspect to mask them from others?

Jimmy Jungle.doc
Deleted in a normal fashion.

SCHEDU~1.EXE
Edited the FAT to change the length of the file.

cover page.jpg
Added extra characters to the filename, edited the FAT so the directory pointed to the wrong starting cluster for the file.

What processes did you (the investigator) use to successfully examine the entire contents of each file?

Jimmy Jungle.doc
Used the DOS undelete command to recover the file.

SCHEDU~1.EXE
Identified the file's boundries by following the zip archive format, then used WinHex to recreate the file.  Alternatively, repaired the file's length in the FAT so the file is available from floppy.  The file was unzipped using WinZip, with the password 'goodtimes' that was found on the disk.

cover page.jpg
Identified the file's boundries by following the JPEG image format, then used WinHex to recreate the file.  Alternatively, repaired the file's starting cluster in the FAT so the file is available from floppy.  Alternatively, used Norton Disk Doctor to recover the file.

*BONUS* What Microsoft program was used to create the Cover Page file? What is your proof (Proof is the key to getting this question right, not just making a guess)?

We believe that Paint 5.1 was used to create the cover page image.  In order to determine the application that created the jpg, the header of the file was disected using WinHex. This deconstruction follows:
FFD8: SOI (Identify a JPEG/JFIF File)
FFE0: JFIF APP0 segment marker
0010: Length (high byte: 00; low byte: 10  -- Length of 16 bytes)
4A46494600: Identifies JFIF
01: Major Revision
01: Minor Revision
01: Units for x/y densities (here, x/y is in dots/inch)
0060: X density (high byte; low byte)
0060: y density (high byte; low byte)
00: thumbnail width
00: thumbnail height
FFDB: Define Quantization Table
0043: Length (high byte; low byte)
00: n bytes QT, n = 64*(precision+1), Precision (0=8bit)

Quantization Table
08060607060508
0707070909080A0C140D0C0B0B0C1912
130F141D1A1F1E1D1A1C1C20242E2720
222C231C1C2837292C30313434341F27
393D38323C2E333432

FFDB: Define another Quantization Table
0043: Length (high byte; low byte)
01: n bytes QT, n = 64*(precision+1), Precision (1=16bit)

Quantization table
0909090C0B0C180D0D1832211C2132323232
32323232323232323232323232323232
32323232323232323232323232323232
3232323232323232323232323232

FFC0: Start of Frame 0
0011: Length
08: Precision (8 bits/sample)
00C7: Height
00D0: Width
03: Number of components
Component #1
01: Y component
22: sampling factor
00: quantization table #
Component #2
02: Cb Component
11: sampling factor
01: quantization table #
Component #3
03: Cr component
11: sampling factor
01: quantization table #

APP0: JFIF segment marker:
~~~~~~~~~~~~~~~~~~~~~~~~~~
  - $ff, $e0 (APP0)
  - length (high byte, low byte), must be >= 16
  - 'JFIF'#0 ($4a, $46, $49, $46, $00), identifies JFIF
  - major revision number, should be 1 (otherwise error)
  - minor revision number, should be 0..2 (otherwise try to decode anyway)
  - units for x/y densities:
     0 = no units, x/y-density specify the aspect ratio instead
     1 = x/y-density are dots/inch
     2 = x/y-density are dots/cm
  - x-density (high byte, low byte), should be <> 0
  - y-density (high byte, low byte), should be <> 0
  - thumbnail width (1 byte)
  - thumbnail height (1 byte)
  - n bytes for thumbnail (RGB 24 bit), n = width*height*3

Table #1. APP0 Segment Information

SOF0: Start Of Frame 0:
~~~~~~~~~~~~~~~~~~~~~~~
  - $ff, $c0 (SOF0)
  - length (high byte, low byte), 8+components*3
  - data precision (1 byte) in bits/sample, usually 8 (12 and 16 not
    supported by most software)
  - image height (2 bytes, Hi-Lo), must be >0 if DNL not supported
  - image width (2 bytes, Hi-Lo), must be >0 if DNL not supported
  - number of components (1 byte), usually 1 = grey scaled, 3 = color YCbCr
    or YIQ, 4 = color CMYK)
  - for each component: 3 bytes
     - component id (1 = Y, 2 = Cb, 3 = Cr, 4 = I, 5 = Q)
     - sampling factors (bit 0-3 vert., 4-7 hor.)
     - quantization table number

Table #2. Frame 0 Segment Information
After testing several Microsoft Image Editing Programs (Image Composer, PhotoDraw 2k and Paint 5.0 and Paint 5.1), it was discovered that, of these, only Microsoft Paint 5.1 (under Windows XP) created a header with an x and y density identical to that of the recovered jpg file (0060 - see above). Therefore, we believe that Paint 5.1 was used to create the cover page image.
Image Composer:    0001,0001
PhotoDraw:    0096,0096
Paint 5.0:    012C,012C
Paint 5.1:    0060,0060
recovered.jpg:    0060,0060


Marc Bayerkohler
Senior Technical Consultant, Schlumberger
marc_hnet@slb.com  

Vince Elliott
Systems Developer, Schlumberger
vince_hnet24@slb.com

Schlumberger