FILE SYSTEM FORENSICS: NTFS MFT, EXT4 INODES, JOURNALING & SLACK CARVING

Architect in-depth file system reconstruction and digital evidence recovery for B.Tech Computer Science & IT: Structural comparative anatomy of Microsoft NTFS and Linux Extended 4 (EXT4); Master File Table (MFT) record headers and standard attribute parsing ($STANDARD_INFORMATION, $FILE_NAME, $DATA); resident vs. non-resident data runs; Alternate Data Streams (ADS); EXT4 Block Groups, Inode tables, directory entries (dir_entry_2), and JBD2 journaling forensics; RAM slack vs. file slack analysis; and automated raw file carving pipelines via header/footer signatures.

πŸ’Ύ On-Disk Structures, Metadata Allocation & Unallocated Residual Recovery

When an adversary, insider threat, or malware deletes a file from an operating system, the underlying data blocks are almost never instantly wiped or overwritten. Instead, modern operating systems modify metadata flags to unlink file references, marking previously allocated clusters as "free" or "unallocated" space available for future write operations. File System Forensics is the technical discipline of parsing low-level on-disk data structures to reconstruct past events, prove data provenance, track anti-forensic deletion attempts, and carve residual artifacts. In the Windows realm, the New Technology File System (NTFS) organizes all volume metadata within the Master File Table ($MFT), recording millisecond-precision timestamps across $STANDARD_INFORMATION and $FILE_NAME attributes while concealing hidden payloads in Alternate Data Streams (ADS). In the Linux domain, the Fourth Extended File System (EXT4) distributes storage across Block Groups, tracking file permissions, extents, and pointer chains inside Inodes while logging atomic transactions in the JBD2 Journal. Forensic investigators exploit these structural realities to recover evidence long after standard OS utilities report it gone.

1. File System Structural Abstraction: NTFS vs. EXT4 Architecture
Architectural schematic contrasting the layout of Windows NTFS against Linux EXT4 on raw storage blocks. Details the master control zones, metadata management structures, allocation units, and journaling models.
NTFS vs. EXT4 On-Disk Partition Layout & Metadata Topography
NTFS VS. EXT4 ON-DISK PARTITION ARCHITECTURE:

1. WINDOWS NTFS (NEW TECHNOLOGY FILE SYSTEM) VOLUME LAYOUT:
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ VBR / Boot Sector  β”‚ Master File Table ($MFT) & System Metadata Filesβ”‚ General Data Cluster   β”‚
β”‚ Sector 0 (512B)    β”‚ $MFT, $LogFile, $Volume, $Bitmap, $BadClust...  β”‚ Storage Area (Files,   β”‚
β”‚ Cluster Size: 4 KB β”‚ Fixed 1024-byte record entries per inode/file   β”‚ Directories, Streams)  β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
 β€’ Core Philosophy : "Everything on an NTFS volume is a file, including metadata itself ($MFT)."
 β€’ Addressing Unit : Logical Cluster Number (LCN) and Virtual Cluster Number (VCN).
 β€’ Transaction Log : $LogFile (Circular transaction logging) & $UsnJrnl (Update Sequence Number).

2. LINUX EXT4 (FOURTH EXTENDED FILE SYSTEM) VOLUME LAYOUT:
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Boot Sector Pad    β”‚ BLOCK GROUP 0               β”‚ BLOCK GROUP 1               β”‚ BLOCK GROUP Nβ”‚
β”‚ 1024 Bytes         β”‚ Superblock | Group Descs... β”‚ Superblock Backup | ...     β”‚ Data Blocks  β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
                       β”‚
                       β–Ό (Internal Architecture of a single EXT4 Block Group)
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Superblock      β”‚ Group Descriptorsβ”‚ Block Bitmap β”‚ Inode Bitmap β”‚ Inode Table  β”‚ Data Blocks  β”‚
β”‚ 1 Block (1024B) β”‚ Block allocationsβ”‚ 1 Block      β”‚ 1 Block      β”‚ Inode list   β”‚ Actual file  β”‚
β”‚ Filesystem info β”‚ per group        β”‚ Free/Used    β”‚ Free/Used    β”‚ (256B each)  β”‚ data clustersβ”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
 β€’ Core Philosophy : Unix inode hierarchy decoupling file names (Directory Entries) from metadata (Inodes).
 β€’ Addressing Unit : Fixed Logical Blocks (typically 4096 bytes).
 β€’ Transaction Log : JBD2 (Journaling Block Device 2 - stored in reserved Inode #8).
Architectural Standard: In NTFS, file names, attributes, timestamps, and data runs are centralized inside the $MFT record entries. In EXT4, the file name is stored independently inside directory data blocks (dir_entry_2), pointing to an inode index number in the Inode Table, which in turn maps data blocks through an extent tree.
Real-Time Example 1: NTFS Master File Table ($MFT) Record Structure & Header Parsing
Analysis of an NTFS $MFT entry. Deconstructs the 1024-byte record layout, the FILE magic identifier, update sequence arrays, record allocation flags, and the variable-length attribute chain.
NTFS 1024-Byte MFT Record Anatomy & Offset Specification
NTFS 1024-BYTE $MFT RECORD ENTRY STRUCTURE:

Offset (Hex)  Length (Bytes)  Field Description
────────────  ──────────────  ─────────────────────────────────────────────────────────────
0x00 - 0x03   4 Bytes         Magic Header Bytes: "FILE" (0x46 0x49 0x4C 0x45) or "BAAD"
0x04 - 0x05   2 Bytes         Offset to Update Sequence Array (Typically 0x0030)
0x06 - 0x07   2 Bytes         Size of Update Sequence Array (S + 1)
0x08 - 0x0F   8 Bytes         $LogFile Sequence Number (LSN)
0x10 - 0x11   2 Bytes         Sequence Number (Increments each time record is reallocated)
0x12 - 0x13   2 Bytes         Hard Link Count (Number of directory references to this file)
0x14 - 0x15   2 Bytes         Offset to First Attribute (Typically 0x0038)
0x16 - 0x17   2 Bytes         Flags: 0x01 = Allocated File | 0x00 = Deleted File
                              Flags: 0x03 = Allocated Directory | 0x02 = Deleted Directory
0x18 - 0x1B   4 Bytes         Used Size of MFT Record (e.g., 0x000001E8 = 488 bytes used)
0x1C - 0x1F   4 Bytes         Allocated Size of MFT Record (Always 0x00000400 = 1024 bytes)
0x20 - 0x27   8 Bytes         File Reference to Base MFT Record (0 if base record)
0x28 - 0x29   2 Bytes         Next Attribute ID
0x38 - End    Variable        VARIABLE ATTRIBUTE CHAIN:
                              β”œβ”€β”€ Attribute 0x10: $STANDARD_INFORMATION (Timestamps, DOS Flags)
                              β”œβ”€β”€ Attribute 0x30: $FILE_NAME (Namespace, Parent Directory Ref)
                              β”œβ”€β”€ Attribute 0x80: $DATA (File Contents / Extent Data Runs)
                              └── End Marker    : 0xFFFFFFFF (End of MFT Record Attributes)
Forensic Significance: The flag at offset 0x16 is the core indicator for deleted file detection. When a user deletes a file, NTFS does not overwrite the record; it flips offset 0x16 from 0x01 (Allocated File) to 0x00 (Deleted File), leaving file names, timestamps, and cluster pointer data runs fully intact for recovery until overwritten.
Real-Time Example 2: NTFS Timestamps: $STANDARD_INFORMATION vs. $FILE_NAME (Timestomping)
Analysis of the dual-timestamp architecture in NTFS. Contrasts the user-accessible $STANDARD_INFORMATION (0x10) timestamps against the kernel-protected $FILE_NAME (0x30) timestamps, detailing how to detect anti-forensic Timestomping attacks.
MACB Timestamps & Timestomp Detection Matrix
NTFS DUAL TIMESTAMP ARCHITECTURE (MACB ARTIFACT MATRIX):

Each MFT record contains TWO independent sets of 64-bit Windows FILETIME timestamps
(Count of 100-nanosecond intervals since January 1, 1601 UTC):

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ TIMESTAMP TYPE   β”‚ $STANDARD_INFORMATION (0x10)        β”‚ $FILE_NAME (0x30)                    β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ Accessibility    β”‚ User Mode Accessible via Win32 API  β”‚ Kernel Mode Protected (System Only)  β”‚
β”‚                  β”‚ SetFileTime() function              β”‚ Updated ONLY on create/rename/move   β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ M: Modified Time β”‚ Last date/time content was altered  β”‚ Last date/time content was altered   β”‚
β”‚ A: Accessed Time β”‚ Last date/time file was read/opened β”‚ Last date/time file was read/opened  β”‚
β”‚ C: MFT Changed   β”‚ Last date/time MFT record modified  β”‚ Last date/time MFT record modified   β”‚
β”‚ B: Birth / Createβ”‚ File creation date/time on volume   β”‚ File creation date/time on volume    β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

TIMESTOMPING EXPLOITATION & FORENSIC TRIAGE:
Adversaries use tools like Meterpreter 'timestomp' to alter $STANDARD_INFORMATION timestamps,
making malicious binaries appear as if installed alongside Windows system files years earlier.

β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ TIMESTOMPED INTRUSION ARTIFACT DETECTED:                                                      β”‚
β”‚ File: C:\Windows\System32\malware.exe                                                         β”‚
β”‚                                                                                               β”‚
β”‚ $STANDARD_INFORMATION (0x10):                                                                 β”‚
β”‚   β€’ Created (B) : 2021-03-15 04:12:00.0000000 UTC (SPOOFED to match Windows OS install)       β”‚
β”‚   β€’ Modified (M): 2021-03-15 04:12:00.0000000 UTC                                             β”‚
β”‚                                                                                               β”‚
β”‚ $FILE_NAME (0x30) - THE IMMUTABLE FORENSIC GROUND TRUTH:                                      β”‚
β”‚   β€’ Created (B) : 2026-11-04 14:22:18.1092841 UTC (TRUE DROP TIME!)                           β”‚
β”‚   β€’ Modified (M): 2026-11-04 14:22:18.1092841 UTC                                             β”‚
β”‚                                                                                               β”‚
β”‚ TRIAGE ANOMALY:                                                                               β”‚
β”‚ 1. $STANDARD_INFO Created (2021) is 5 years EARLIER than $FILE_NAME Created (2026)!           β”‚
β”‚ 2. Sub-second precision zeroed out: 00.0000000 (Human APIs zero nanoseconds; kernel does not).β”‚
β”‚ CONCLUSION: Confirmed Timestomping! True malicious file drop occurred on Nov 4, 2026.         β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
Investigative Invariant: Standard user-space Windows APIs (including SetFileTime) can only update the $STANDARD_INFORMATION attribute. The $FILE_NAME attribute is updated exclusively by the NTFS kernel driver during file creation, rename, or directory movement operations. Comparing 0x10 and 0x30 reveals timestomping attempts.
Real-Time Example 3: Resident vs. Non-Resident Data Runs & Alternate Data Streams (ADS)
Analysis of the NTFS $DATA (0x80) attribute. Contrasts Resident data (payload embedded directly inside the 1024-byte MFT record) with Non-Resident data runs, and illustrates how adversaries conceal tools inside Alternate Data Streams (payload.exe:hidden.vbs).
NTFS Resident Attributes, Data Run Decompression & Alternate Data Streams
NTFS $DATA ATTRIBUTE (0x80) ALLOCATION ARCHITECTURE:

1. RESIDENT DATA (FILES SMALLER THAN ~700 BYTES):
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ β€’ If file content fits inside the available space of the 1024-byte MFT record:                β”‚
β”‚ β€’ Non-Resident Flag at offset 0x08 = 0x00 (Resident).                                         β”‚
β”‚ β€’ Content is stored directly within the MFT record itself! Zero external clusters allocated.  β”‚
β”‚ β€’ Forensic Value: Even if the entire disk clusters are wiped, the file content remains fully  β”‚
β”‚   preserved inside the 1024-byte MFT entry!                                                   β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

2. NON-RESIDENT DATA RUNS (FILES LARGER THAN ~700 BYTES):
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ β€’ Non-Resident Flag at offset 0x08 = 0x01 (Non-Resident).                                     β”‚
β”‚ β€’ Data is stored in external clusters. The MFT stores compressed "Data Runs":                 β”‚
β”‚                                                                                               β”‚
β”‚ Raw Data Run Hex Example: 0x32 0x01 0x20 0x05 0x40 0x12                                       β”‚
β”‚   β”œβ”€β”€ Run Header: 0x32 (High nibble = 3 bytes offset | Low nibble = 2 bytes length)           β”‚
β”‚   β”œβ”€β”€ Length    : 0x0120 = 288 clusters allocated to this run                                 β”‚
β”‚   └── Offset    : 0x124005 = Starting at Logical Cluster Number (LCN) 1,196,037               β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

3. ALTERNATE DATA STREAMS (ADS) CONCEALMENT:
NTFS supports multiple data streams attached to a single file entry:
  β€’ Default stream: file.txt (Uniquely referenced as file.txt:$DATA)
  β€’ Hidden stream : file.txt:malware.exe (Concealed payload stream)

Command-Line Creation & Forensic Extraction:
$ type payload.exe > "C:\Users\Public\readme.txt:beacon.exe"
$ dir /r "C:\Users\Public\"
  readme.txt                  15 bytes
  readme.txt:beacon.exe:$DATA 4,194,304 bytes  <-- Large hidden stream revealed!
Execution bypass:
$ wmic process call create "C:\Users\Public\readme.txt:beacon.exe"
ADS Forensic Triage: Standard Windows Explorer and default 'dir' commands display only the primary file size (15 bytes), completely hiding the 4MB executable in the stream. Investigators detect Alternate Data Streams using 'dir /r', Sysinternals 'streams -s', or by parsing 0x80 attribute entries named with non-empty string headers in the $MFT.
Real-Time Example 4: EXT4 File System Architecture: Block Groups, Inodes & Extents
Analysis of the Linux EXT4 file system. Details Block Group descriptors, Inode table structures (256-byte records), mode flags, and modern Extent Trees (replacing indirect block pointers) for tracking fragmented data allocations.
Linux EXT4 Inode Data Structure (256 Bytes) & Extent Tree Geometry
LINUX EXT4 INODE STRUCTURE (struct ext4_inode - 256 BYTES):

Offset (Hex)  Length (Bytes)  Field Description & Forensic Role
────────────  ──────────────  ─────────────────────────────────────────────────────────────
0x00 - 0x01   2 Bytes         i_mode: File Type (Regular, Dir, Link) & POSIX Permissions
0x02 - 0x03   2 Bytes         i_uid: Owner User ID (Lower 16 bits)
0x04 - 0x07   4 Bytes         i_size_lo: File Size in bytes (Lower 32 bits)
0x08 - 0x0B   4 Bytes         i_atime: Last Access Time (Unix Epoch 32-bit timestamp)
0x0C - 0x0F   4 Bytes         i_ctime: Inode Status Change Time (Metadata change)
0x10 - 0x13   4 Bytes         i_mtime: Last Data Modification Time
0x14 - 0x17   4 Bytes         i_dtime: Deletion Time (CRITICAL: 0 if allocated; >0 if DELETED!)
0x18 - 0x19   2 Bytes         i_gid: Group ID (Lower 16 bits)
0x1A - 0x1B   2 Bytes         i_links_count: Hard Link Count (When reaches 0, blocks unlinked)
0x1C - 0x1F   4 Bytes         i_blocks_lo: 512-byte block count allocated to file
0x20 - 0x23   4 Bytes         i_flags: File flags (e.g. 0x80000 = EXT4_EXTENTS_FL)
0x28 - 0x63   60 Bytes        i_block / EXT4 EXTENT TREE:
                              Maps file offsets directly to physical disk blocks!

EXT4 EXTENT TREE ARCHITECTURE (Up to 4 extents fit directly inside the inode!):
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Extent Header (12 Bytes)β”‚ eh_magic: 0xF30A (Magic)     β”‚ eh_entries: Count of extents β”‚
β”‚                         β”‚ eh_max: Max capacity (4)     β”‚ eh_depth: 0 = Leaf Node      β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ Extent Leaf 1 (12 Bytes)β”‚ ee_block: Logical block 0    β”‚ ee_len: 128 contiguous blocksβ”‚
β”‚                         β”‚ ee_start: Physical Block # 4,194,304 on disk partition      β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
EXT4 Deletion Behavior: When a file is deleted in EXT4: 1. The kernel records the Unix epoch timestamp into 'i_dtime' (Deletion Time). 2. The hard link counter 'i_links_count' drops to 0. 3. The directory entry in the parent directory has its inode pointer zeroed (or skipped over). 4. In modern EXT4, the extent tree (i_block) is often cleared to prevent filesystem corruption during asynchronous allocation, making unallocated block carving necessary for deleted recovery.
Real-Time Example 5: Linux JBD2 Journaling Forensics: Carving Deleted Evidence from Journal Logs
Analysis of the Linux Journaling Block Device 2 (JBD2). Explains journaling modes (journal, ordered, writeback), descriptor blocks, commit transactions, and recovering deleted inodes from /journal (Inode #8) via ext4magic.
EXT4 JBD2 Journaling Ring-Buffer & Transaction Commit Lifecycle
EXT4 JOURNALING BLOCK DEVICE (JBD2) FORENSIC TRANSACTION FLOW:

THE THREE EXT4 JOURNALING MODES:
1. data=journal   : Full logging. Both file metadata AND data contents are written to journal
                    before being flushed to disk. Highest forensic recoverable artifact yield!
2. data=ordered   : Default mode. File data is written to disk first; then metadata is logged
                    to journal. High metadata recovery rate.
3. data=writeback : Fast mode. Metadata changes are journaled, but data write order is not enforced.

JBD2 CIRCULAR RING-BUFFER ARCHITECTURE (INODE #8):
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ JOURNAL SUPERBLOCK (Magic: 0xC03B3998 | Block Size: 4096B | First Transaction Block: 1)       β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ TRANSACTION N:                                                                                β”‚
β”‚  β”œβ”€β”€ DESCRIPTOR BLOCK: Lists the filesystem blocks modified in this transaction.              β”‚
β”‚  β”œβ”€β”€ METADATA COPY   : Contains snapshot of INODE and DIRECTORY ENTRY before deletion!        β”‚
β”‚  └── COMMIT BLOCK    : Invariant timestamp marking successful atomic commit.                  β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ TRANSACTION N+1 (The Deletion Transaction):                                                   β”‚
β”‚  β”œβ”€β”€ Logs updated inode with i_dtime populated and links_count = 0.                           β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

HOW FORENSIC TOOLS (ext4magic) RECOVER DELETED FILES FROM THE JOURNAL:
Even if an attacker wipes an EXT4 inode and overwrites directory entries, the prior state of the
inode exists preserved inside the JBD2 circular journal ring-buffer for hours or days!

Recovery Command:
$ sudo ext4magic /dev/sdb1 -r -f /var/log/auth.log -d /recovered_evidence/
[+] Scans JBD2 journal ring-buffer.
[+] Reconstructs auth.log inode state prior to deletion transaction.
[+] Successfully restores deleted log file with original timestamps intact!
Investigative Value: The JBD2 journal is a chronological record of file operations. In anti-forensics incidents where an attacker removes files (rm -rf /var/log/*), inspecting the journal via ext4magic or debugfs allows examiners to extract the deleted inodes and their original timestamps from prior transactions.
Real-Time Example 6: Residual Space Forensics: RAM Slack vs. File Slack Analysis
Analysis of residual storage space. Deconstructs the anatomical boundary between logical file sizes and physical cluster allocations, contrasting RAM Slack (sector padding) with Drive/File Slack (cluster padding) as covert data exfiltration channels.
RAM Slack vs. Drive Slack Physical Cluster Boundary Geometry
RESIDUAL SPACE ARCHITECTURE: RAM SLACK VS. FILE / DRIVE SLACK:

Storage Scenario:
β€’ Filesystem Cluster Size = 4096 Bytes (8 sectors of 512 bytes each).
β€’ Suspect writes a secret message: file.txt = 600 Bytes total size.

PHYSICAL ALLOCATION OF A 4096-BYTE CLUSTER ON DISK:
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ SECTOR 1 (512 Bytes)   β”‚ SECTOR 2 (512 B)  β”‚ SECTORS 3 THROUGH 8 (3072 Bytes)                 β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ [Logical File Content] β”‚ [Logical File]    β”‚ [UNALLOCATED SECTORS OF THE ALLOCATED CLUSTER]   β”‚
β”‚ 512 Bytes of file.txt  β”‚ 88 Bytes of file  β”‚ Zero bytes of file.txt belong here!              β”‚
β”‚                        β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚                        β”‚ [RAM SLACK]       β”‚ [DRIVE SLACK / FILE SLACK]                       β”‚
β”‚                        β”‚ 424 Bytes padding β”‚ 3072 Bytes of residual unallocated sector data!  β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜

DEFINITIONS & FORENSIC CHARACTERISTICS:
1. RAM SLACK:
   β€’ The space from the logical end of the file (Byte 600) to the end of that specific SECTOR (Byte 1024).
   β€’ Size: 512 - (600 - 512) = 424 Bytes.
   β€’ History: In older Windows OSs (Windows 95/98), the OS dumped random memory buffer bytes to pad the
     sector, leaking passwords, network packets, and RAM fragments!
   β€’ Modern OSs (NTFS on Windows 10/11, Linux): Zero-pads RAM slack with NULL bytes (\x00) for security.

2. DRIVE SLACK (FILE SLACK / CLUSTER SLACK):
   β€’ The remaining full sectors from the end of the RAM slack sector to the end of the CLUSTER.
   β€’ Sectors 3 through 8 = 6 * 512 = 3072 Bytes.
   β€’ CRITICAL FORENSIC REALITY: Modern operating systems NEVER zero out drive slack!
   β€’ Drive slack retains whatever raw data previously occupied those physical sectors before the new
     small file was written!
   β€’ Anti-Forensics Abuse: Attackers use tools like 'bmap' to hide stolen cryptographic keys or
     passwords directly inside the drive slack of benign system files!
Carving Invariant: Forensic tools like The Sleuth Kit (TSK) use utilities like 'blkcalc' and 'dls' to extract data residing in slack space. When an organization overwrites an old confidential document with a new small document, the drive slack of the cluster often contains recoverable paragraphs of the deleted predecessor.
Real-Time Example 7: Automated Raw File Carving: Magic Numbers & Header/Footer Parsing in Python
Building an automated raw data carving tool in Python from first principles. Scans unallocated disk images for file signatures (magic bytes for JPEG, PNG, PDF), resolves boundaries, and reconstructs orphaned files without relying on file system metadata.
Python (Automated Raw Signature File Carver: carve_signatures.py)
import os
import sys

class RawSignatureCarver:
    """
    Automated File Carving Engine:
    Carves files from raw, unallocated disk images based on magic byte headers and footers.
    Operates independently of corrupted or wiped MFT/Inode metadata structures.
    """
    SIGNATURE_DATABASE = {
        "JPEG": {
            "header": b"\xFF\xD8\xFF\xE0",
            "footer": b"\xFF\xD9",
            "max_size": 20 * 1024 * 1024 # 20 MB max
        },
        "PNG": {
            "header": b"\x89\x50\x4E\x47\x0D\x0A\x1A\x0A",
            "footer": b"\x49\x45\x4E\x44\xAE\x42\x60\x82",
            "max_size": 20 * 1024 * 1024
        },
        "PDF": {
            "header": b"%PDF-",
            "footer": b"%%EOF",
            "max_size": 50 * 1024 * 1024 # 50 MB max
        }
    }

    def __init__(self, disk_image_path: str, output_directory: str):
        self.image_path = disk_image_path
        self.output_dir = output_directory
        os.makedirs(self.output_dir, exist_ok=True)

    def execute_carve(self):
        print(f"[*] Commencing signature carving on: {self.image_path}...")
        with open(self.image_path, "rb") as f:
            data = f.read()

        total_bytes = len(data)
        carved_manifest = []

        for ftype, sigs in self.SIGNATURE_DATABASE.items():
            header = sigs["header"]
            footer = sigs["footer"]
            max_len = sigs["max_size"]
            start_pos = 0

            while True:
                # Locate header signature in raw byte stream
                header_idx = data.find(header, start_pos)
                if header_idx == -1:
                    break # No more matching headers found

                # Search for matching footer signature within max_size boundary
                search_limit = min(header_idx + max_len, total_bytes)
                footer_idx = data.find(footer, header_idx + len(header), search_limit)

                if footer_idx != -1:
                    end_idx = footer_idx + len(footer)
                    extracted_file = data[header_idx:end_idx]

                    file_name = f"carved_{header_idx}_{ftype.lower()}.{ftype.lower()}"
                    save_path = os.path.join(self.output_dir, file_name)

                    with open(save_path, "wb") as out:
                        out.write(extracted_file)

                    carved_manifest.append({
                        "file_type": ftype,
                        "offset_hex": f"0x{header_idx:08X}",
                        "size_bytes": len(extracted_file),
                        "saved_file": file_name
                    })

                    # Advance search position past current carved file
                    start_pos = end_idx
                else:
                    # Footer not found; advance to next byte to avoid infinite loop
                    start_pos = header_idx + 1

        return carved_manifest

if __name__ == "__main__":
    # Demonstration execution with simulated unallocated block containing mock PDF
    MOCK_DISK = "unallocated_block.raw"
    with open(MOCK_DISK, "wb") as f:
        f.write(b"\x00" * 4096) # Initial slack padding
        f.write(b"%PDF-1.7\n1 0 obj<<>>endobj\ntrailer<<>>\n%%EOF") # Embedded PDF file
        f.write(b"\x00" * 4096) # Trailing slack padding

    carver = RawSignatureCarver(MOCK_DISK, "./carved_output")
    print("==========================================================================")
    print("      AUTOMATED RAW SIGNATURE FILE CARVING ENGINE REPORT")
    print("==========================================================================")
    results = carver.execute_carve()

    for item in results:
        print(f"[βœ”] CARVED {item['file_type']:<5} | Offset: {item['offset_hex']} | Size: {item['size_bytes']} bytes βž” {item['saved_file']}")
    
    # Cleanup mock image
    if os.path.exists(MOCK_DISK):
        os.remove(MOCK_DISK)
Carving Telemetry Output: ========================================================================== AUTOMATED RAW SIGNATURE FILE CARVING ENGINE REPORT ========================================================================== [*] Commencing signature carving on: unallocated_block.raw... [βœ”] CARVED PDF | Offset: 0x00001000 | Size: 44 bytes βž” carved_4096_pdf.pdf Carving Theory: File carving scans continuous byte arrays for magic numbers (e.g. %PDF- and %%EOF) without consulting file system tables. Its primary limitation is file fragmentation: if a file is split across non-contiguous clusters on disk, simple linear header-to-footer carving incorporates interleaved data from unrelated files, producing corrupted images.
Real-Time Example 8: Debugging File System Forensic Pitfalls & Triage
Troubleshooting common file system forensic analysis mistakes: SSD TRIM destruction of deleted evidence, interpreting MFT update sequence array fixes, and handling time-zone skew across FAT32 and NTFS boundaries.
File System Forensic Pitfalls & Analytical Countermeasures
FILE SYSTEM FORENSIC DEFECTS & ANALYTICAL TRIAGE:

// ==============================================================================
// PITFALL 1: SOLID STATE DRIVE (SSD) TRIM COMMAND PURGING EVIDENCE
// SYMPTOM: An investigator images an SSD 4 hours after a user deleted critical files.
// When carving unallocated space, ALL unallocated sectors return pure 0x00 bytes!
// CAUSE: Active SSD TRIM (ATA Data Set Management command).
// Unlike magnetic HDDs (where data persists until overwritten), SSDs run background
// Garbage Collection to maintain NAND write speeds, immediately zeroing unlinked blocks!
// REMEDIATION:
// 1. In dynamic incidents on SSDs, disconnect physical power immediately if the machine
//    is off to prevent scheduled idle TRIM cleanup routines!
// 2. Focus analysis on the MFT, Inode tables, and JBD2 journal, which often retain
//    resident file content and metadata records even when external extents are trimmed.
// ==============================================================================

// ==============================================================================
// PITFALL 2: IGNORING UPDATE SEQUENCE NUMBER (USN) ARRAY FIXUPS IN MFT RECORDS
// DEFECT: Analyst reads raw MFT record bytes and notices the final 2 bytes of every
// 512-byte sector do not match expected data runs.
// CAUSE: NTFS Update Sequence Array (USA) protection against partial sector writes.
// When writing an MFT record to disk, NTFS swaps out the last 2 bytes of each 512-byte
// sector and stores them in the USN array at offset 0x30, replacing the sector bytes
// with a 2-byte update sequence signature.
// REMEDIATION: The forensic parser must apply "fixup processing": read the genuine
// bytes out of the USN array and restore them into the sector ends before parsing attributes!
// ==============================================================================

// ==============================================================================
// PITFALL 3: FAT32 VS. NTFS TIME ZONE INTERPRETATION (THE LOCAL TIME CONFUSION)
// DEFECT: An investigator compares timestamps from a USB flash drive (FAT32) against
// a system disk (NTFS), reporting a 5-hour 30-minute discrepancy as an anomaly.
// CAUSE: NTFS stores FILETIME in strict UTC (Coordinated Universal Time).
// FAT32 stores file timestamps in LOCAL BIOS TIME with ZERO time zone metadata!
// REMEDIATION: Always determine the underlying file system type before correlating
// timelines, and normalize FAT32 timestamps to UTC based on the suspect system's active
// time zone registry key (HKLM\SYSTEM\CurrentControlSet\Control\TimeZoneInformation).
// ==============================================================================
Triage Summary: - SSD TRIM Erasure: Shifted triage focus from carved data blocks to resident MFT attributes and journal transaction records. - USN Array Disruption: Implemented Update Sequence Array fixups before evaluating attribute headers. - Timestamp Discrepancies: Normalized local FAT32 USB timestamps against NTFS UTC timestamps using system registry time zone values.
Real-Time Example 9: File System Forensics Specification Reference Matrix
A comprehensive reference summary pairing file systems, metadata structures, evidence locations, operational tools, and academic exam keywords for B.Tech engineering students.
File System Forensics Specification Reference Matrix
FILE SYSTEM FORENSICS SPECIFICATION MATRIX:
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ FILE SYSTEM / ARTIFACT  β”‚ STRUCTURAL LOCATION          β”‚ RECORD SIZE      β”‚ FORENSIC EVIDENCE YIELD          β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ NTFS $MFT               β”‚ \$MFT (Root of partition)    β”‚ 1024 Bytes       β”‚ Centralized metadata ledger;     β”‚
β”‚ (Master File Table)     β”‚ Fixed sequential record arrayβ”‚ per entry        β”‚ file names, sizes, data runs     β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ $STANDARD_INFORMATION   β”‚ MFT Attribute 0x10           β”‚ Variable         β”‚ Primary MACB timestamps,         β”‚
β”‚                         β”‚ (User mode accessible)       β”‚ (~72 bytes)      β”‚ DOS attributes; reveals timestompβ”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ $FILE_NAME              β”‚ MFT Attribute 0x30           β”‚ Variable         β”‚ Kernel protected MACB times,     β”‚
β”‚                         β”‚ (Kernel mode protected)      β”‚ (~100 bytes)     β”‚ file names up to 255 chars       β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ Alternate Data Streams  β”‚ MFT Attribute 0x80           β”‚ Variable         β”‚ Hidden payloads, Zone.Identi-    β”‚
β”‚ (NTFS ADS)              β”‚ with non-empty name string   β”‚                  β”‚ fier tracking internet drops     β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ EXT4 Inode Table        β”‚ Inode Table blocks per group β”‚ 256 Bytes        β”‚ POSIX permissions, UID/GID,      β”‚
β”‚                         β”‚ (Allocated in block groups)  β”‚ per inode        β”‚ extent tree, deletion time       β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ EXT4 Directory Entry    β”‚ Directory data blocks        β”‚ Variable         β”‚ Unlinks file name from inode;    β”‚
β”‚ (dir_entry_2)           β”‚ inode #, rec_len, name_len   β”‚                  β”‚ reveals deleted file names       β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ JBD2 Journal            β”‚ Inode #8 (/journal)          β”‚ Circular         β”‚ Pre-deletion inode snapshots;    β”‚
β”‚ (EXT4 Journaling Engine)β”‚ Ring-buffer data blocks      β”‚ ring-buffer      β”‚ transaction timeline metadata    β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ Drive / File Slack      β”‚ Residual space between file  β”‚ Variable         β”‚ Data remnants from deleted       β”‚
β”‚                         β”‚ end and end of cluster       β”‚ (< Cluster size) β”‚ predecessor files; hidden data   β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ File Carving            β”‚ Unallocated cluster blocks   β”‚ Independent of   β”‚ Reconstructs fragmented or       β”‚
β”‚ (Signature Carving)     β”‚ Magic headers and footers    β”‚ file system      β”‚ unlinked files without tables    β”‚
β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€
β”‚ TOP 5 FILE SYSTEM FORENSIC BEST PRACTICES:                                                                   β”‚
β”‚ 1. Always Check Dual Timestamps: Compare 0x10 ($STANDARD_INFO) and 0x30 ($FILE_NAME) for timestomping.       β”‚
β”‚ 2. Audit Alternate Data Streams: Search for hidden streams on suspected download drop directories.           β”‚
β”‚ 3. Recover Journal Records Early: Mine JBD2 journal buffers before circular transactions overwrite them      β”‚
β”‚ 4. Account for SSD TRIM: Preserve volatile and metadata structures immediately on flash media.               β”‚
β”‚ 5. Parse Slack Space: Carve cluster slack space to recover residual fragments of prior documents.            β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
Academic & Professional Standard: File system forensics provides the physical foundation for reconstructing cybercrime events. Mastering the internal mechanics of the NTFS Master File Table, EXT4 inode extent trees, JBD2 journaling buffers, and residual slack carving equips B.Tech students to analyze damaged partitions, recover deleted artifacts, and build defensible digital forensic timelines.