Zum Inhalt springen
1337skills 1337skills - Referenzen

SRUM-DUMP Cheat Sheet

Overview

SRUM-DUMP is a forensic tool that extracts and parses data from the Windows System Resource Usage Monitor (SRUM) database. SRUM is a Windows feature introduced in Windows 8 that tracks detailed application resource usage including network data sent/received per application, application execution history with timestamps, energy usage statistics, push notifications, and Windows Timeline activity. This data is stored in an ESE (Extensible Storage Engine) database at C:\Windows\System32\sru\SRUDB.dat and retains up to 30-60 days of historical data.

For digital forensics and incident response, SRUM is invaluable because it provides evidence of program execution, network activity per process, and data transfer volumes that persist even after the processes have terminated and event logs have rolled over. SRUM data can prove that a specific application ran at a particular time, how much data it transferred over the network, and which network interfaces were used — critical evidence for establishing timelines of attacker activity, data exfiltration, and malware execution. The SRUM database survives many anti-forensic techniques since it is managed by the operating system and not easily tampered with.

Installation

SRUM-DUMP (Python)

# Install SRUM-DUMP
pip install srum-dump

# Or clone from GitHub
git clone https://github.com/MarkBaggett/srum-dump.git
cd srum-dump
pip install -r requirements.txt

# Verify
python srum_dump.py --help

SrumECmd (Eric Zimmerman’s Tools)

# Download SrumECmd from Eric Zimmerman's tools
# https://ericzimmerman.github.io/#!index.md

# Download via script
Invoke-WebRequest -Uri "https://f001.backblazeb2.com/file/EricZimmermanTools/net6/SrumECmd.zip" -OutFile SrumECmd.zip
Expand-Archive SrumECmd.zip -DestinationPath C:\Tools\SrumECmd

# Verify
C:\Tools\SrumECmd\SrumECmd.exe --help

Prerequisites

# SRUM-DUMP requires the SRUM database and SOFTWARE registry hive
# Live system locations:
#   Database: C:\Windows\System32\sru\SRUDB.dat
#   Registry: C:\Windows\System32\config\SOFTWARE

# For live extraction, stop the diagnostic service first:
net stop "Diagnostic Policy Service"

# Copy the database
copy C:\Windows\System32\sru\SRUDB.dat C:\Evidence\SRUDB.dat
copy C:\Windows\System32\config\SOFTWARE C:\Evidence\SOFTWARE

# Restart the service
net start "Diagnostic Policy Service"

Core Commands

SRUM-DUMP

# Basic extraction to Excel
python srum_dump.py -i SRUDB.dat -t SRUM_TEMPLATE2.xlsx -o output.xlsx

# With SOFTWARE hive for SID resolution
python srum_dump.py -i SRUDB.dat -t SRUM_TEMPLATE2.xlsx -r SOFTWARE -o output.xlsx

# Specify output directory
python srum_dump.py -i SRUDB.dat -t SRUM_TEMPLATE2.xlsx -o C:\Analysis\srum_output.xlsx

SrumECmd

# Parse SRUM database to CSV
SrumECmd.exe -f C:\Evidence\SRUDB.dat -r C:\Evidence\SOFTWARE --csv C:\Analysis\

# Parse with all output formats
SrumECmd.exe -f C:\Evidence\SRUDB.dat -r C:\Evidence\SOFTWARE --csv C:\Analysis\ --debug

# Parse from mounted image
SrumECmd.exe -f E:\Windows\System32\sru\SRUDB.dat -r E:\Windows\System32\config\SOFTWARE --csv C:\Analysis\

# Live system parsing (requires stopping DPS)
SrumECmd.exe -f C:\Windows\System32\sru\SRUDB.dat -r C:\Windows\System32\config\SOFTWARE --csv C:\Analysis\

SRUM Data Tables

Key Database Tables

TableGUIDDescription
Application Resource Usage{D10CA2FE-6FCF-4F6D-848E-B2E99266FA89}App execution with CPU, memory, I/O
Network Data Usage{973F5D5C-1D90-4944-BE8E-24B94231A174}Network bytes sent/received per app
Network Connectivity{DD6636C4-8929-4683-974E-22C046A43763}Network interface connectivity data
Energy Usage{FEE4E14F-02A9-4550-B5CE-5FA2DA202E37}Application energy consumption
Windows Push Notifications{D10CA2FE-6FCF-4F6D-848E-B2E99266FA86}Push notification data
App Timeline Provider{5C8CF1C7-7257-4F13-B223-970EF5939312}Windows Timeline activity

Application Resource Usage Fields

FieldDescription
TimeStampHour-granularity timestamp
AppIdApplication identifier (EXE path or SID)
UserIdUser SID who ran the application
ForegroundCycleTimeCPU cycles in foreground
BackgroundCycleTimeCPU cycles in background
FaceTimeTime in foreground (100ns intervals)
ForegroundBytesReadDisk bytes read in foreground
ForegroundBytesWrittenDisk bytes written in foreground
ForegroundNumReadOperationsCount of read operations
ForegroundNumWriteOperationsCount of write operations

Network Data Usage Fields

FieldDescription
TimeStampHour-granularity timestamp
AppIdApplication identifier
UserIdUser SID
InterfaceLuidNetwork interface identifier
L2ProfileIdNetwork profile (connected network)
BytesSentBytes sent over network
BytesRecvdBytes received from network

Analysis Techniques

Identifying Program Execution

# Parse SRUM output CSV for specific application
# After running SrumECmd:
Import-Csv "C:\Analysis\SrumECmd_AppResourceUsage.csv" |
    Where-Object { $_.ExeInfo -match "powershell|cmd|wscript|cscript" } |
    Sort-Object Timestamp |
    Select-Object Timestamp, ExeInfo, UserId, ForegroundCycleTime |
    Format-Table

# Find applications that transferred significant data
Import-Csv "C:\Analysis\SrumECmd_NetworkUsages.csv" |
    Where-Object { [int64]$_.BytesSent -gt 10MB } |
    Sort-Object { [int64]$_.BytesSent } -Descending |
    Select-Object Timestamp, ExeInfo, BytesSent, BytesRecvd |
    Format-Table

Detecting Data Exfiltration

# Find applications with high outbound data transfer
Import-Csv "C:\Analysis\SrumECmd_NetworkUsages.csv" |
    Group-Object ExeInfo |
    ForEach-Object {
        $totalSent = ($_.Group | Measure-Object -Property BytesSent -Sum).Sum
        $totalRecvd = ($_.Group | Measure-Object -Property BytesRecvd -Sum).Sum
        [PSCustomObject]@{
            Application = $_.Name
            TotalSentMB = [math]::Round($totalSent / 1MB, 2)
            TotalRecvdMB = [math]::Round($totalRecvd / 1MB, 2)
            Ratio = if ($totalRecvd -gt 0) { [math]::Round($totalSent / $totalRecvd, 2) } else { "N/A" }
        }
    } |
    Sort-Object TotalSentMB -Descending |
    Format-Table

# Timeline of network activity for suspicious application
Import-Csv "C:\Analysis\SrumECmd_NetworkUsages.csv" |
    Where-Object { $_.ExeInfo -match "suspicious_app" } |
    Sort-Object Timestamp |
    Select-Object Timestamp, BytesSent, BytesRecvd, InterfaceLuid |
    Format-Table

Correlating with Other Artifacts

# Combine SRUM with Prefetch data
# SRUM shows WHEN and HOW MUCH data was transferred
# Prefetch shows execution count and files accessed

# Cross-reference SRUM app execution with event logs
$srumApps = Import-Csv "C:\Analysis\SrumECmd_AppResourceUsage.csv" |
    Select-Object -ExpandProperty ExeInfo -Unique

$eventLogApps = Get-WinEvent -FilterHashtable @{
    LogName='Microsoft-Windows-Sysmon/Operational'
    Id=1
} | Select-Object -ExpandProperty Properties | 
    Select-Object -First 1 -ExpandProperty Value

# Find apps in SRUM but not in event logs (potential log evasion)
$srumApps | Where-Object { $_ -notin $eventLogApps }

Advanced Usage

Forensic Image Analysis

# Mount forensic image
# (Using Arsenal Image Mounter, FTK Imager, or similar)

# Extract SRUM from mounted image
$imageDrive = "E:"
$evidence = "C:\Evidence\Case001"
New-Item -ItemType Directory -Force -Path $evidence

Copy-Item "$imageDrive\Windows\System32\sru\SRUDB.dat" "$evidence\"
Copy-Item "$imageDrive\Windows\System32\config\SOFTWARE" "$evidence\"

# Parse with SrumECmd
SrumECmd.exe -f "$evidence\SRUDB.dat" -r "$evidence\SOFTWARE" --csv "$evidence\parsed\"

Automated Reporting

#!/usr/bin/env python3
"""Generate SRUM analysis report."""
import csv
import json
from collections import defaultdict
from datetime import datetime

def analyze_srum(network_csv, app_csv):
    """Analyze SRUM data for suspicious activity."""
    findings = {
        'high_data_transfer': [],
        'unusual_hours': [],
        'suspicious_apps': []
    }

    # Analyze network usage
    with open(network_csv, 'r') as f:
        reader = csv.DictReader(f)
        app_totals = defaultdict(lambda: {'sent': 0, 'recv': 0})
        for row in reader:
            app = row.get('ExeInfo', 'Unknown')
            sent = int(row.get('BytesSent', 0))
            recv = int(row.get('BytesRecvd', 0))
            app_totals[app]['sent'] += sent
            app_totals[app]['recv'] += recv

    # Flag high data transfer
    for app, totals in app_totals.items():
        if totals['sent'] > 100 * 1024 * 1024:  # > 100MB sent
            findings['high_data_transfer'].append({
                'app': app,
                'sent_mb': round(totals['sent'] / (1024*1024), 2),
                'recv_mb': round(totals['recv'] / (1024*1024), 2)
            })

    return findings

results = analyze_srum('NetworkUsages.csv', 'AppResourceUsage.csv')
print(json.dumps(results, indent=2))

Timeline Generation

# Create unified timeline from SRUM data
Import-Csv "C:\Analysis\SrumECmd_AppResourceUsage.csv" |
    Select-Object @{N='Timestamp';E={$_.Timestamp}},
                  @{N='Type';E={'AppExecution'}},
                  @{N='Detail';E={$_.ExeInfo}},
                  @{N='User';E={$_.UserId}} |
    Export-Csv "C:\Analysis\srum_timeline.csv" -NoTypeInformation

Import-Csv "C:\Analysis\SrumECmd_NetworkUsages.csv" |
    Select-Object @{N='Timestamp';E={$_.Timestamp}},
                  @{N='Type';E={'NetworkUsage'}},
                  @{N='Detail';E={"$($_.ExeInfo) - Sent: $($_.BytesSent) Recv: $($_.BytesRecvd)"}},
                  @{N='User';E={$_.UserId}} |
    Export-Csv "C:\Analysis\srum_timeline.csv" -NoTypeInformation -Append

Troubleshooting

IssueSolution
Database lockedStop “Diagnostic Policy Service” before copying SRUDB.dat
Corrupted databaseUse esentutl /r sru /d to attempt recovery on a copy
Missing SOFTWARE hiveTool still works but SIDs won’t resolve to usernames
Empty resultsVerify SRUDB.dat is from Windows 8+ and contains data
Permission deniedRun as Administrator or extract from forensic image
Large output fileFilter by date range or specific applications during analysis
SID resolution failsEnsure correct SOFTWARE hive matches the SRUM database source
Timestamp interpretationSRUM timestamps are in UTC; convert to local time for analysis