The MacScavenger is a system built to circumvent such MAC address randomization by applying signal strength-based localization to distinguish different devices otherwise non-distinguishable by looking at the information they contain.

It consists of four main components:

• The Shell: Responsible for user interaction
• The Monitors: Responsible for data collection
• The Sync: Responsible for interacting and synchronizing the monitor nodes and store the data.
• The Analyzer: Responsible for interpreting the collected data

The user interacts with the system by starting the Shell via python MacScavengerShell.py. From there the different possible commands are listed typing help.

The two main uses of the MacScavenger are Data Collection and Data Analysis:

To gather data, the code for the monitor nodes must be started via python MacScavengerMonitor.py PORT_NUMBER on the devices responsible for data collection. After successfully deploying the monitor nodes, the Shell is used to start and stop the data collection process by tying setup, start or stop.

To configure a valid database source for storing the monitored data. A folder must be placed in the folder SyncDataBaseInterfaces, implementing the interface SyncDataBaseBaseClass. This class can then be loaded in the MacScavengerSync class and assigned as database with the following code:
self.database = CustomDataBase().

The data anlysis process is started via the shell by typing analyze into the shell.

The analysis process can be tested by using the total_data.json file in the folder json_data:

Welcome! Type ? to list commands
MacScavenger> analyze

Please enter Access Point Positons in the following form {"ap1":(0,0), "ap2":(5,0),"ap3":(5,5),"ap4":(0,5)}
{"tinkerboard1":(0,0),"tinkerboard2":(2,1),"tinkerboard3":(2,5),"tinkerboard4":(0,0)}

Analyzer has the following Configuration:
- Interval Size in seconds: 10
- Assumed Walking Speed in km/h: 2
- In Burst Time Threshold in seconds: 1
- Min. AP detection Rate: 3
- Verbosity: 0
Do you agree with this configuration? Type y/n!
y

Please enter a valid data path
json_data/total_data.json

The specified data source is a valid file
Starting Analysis ...

Summary:                                                                             
Approximately 425 different recognizable devices on site that were detected by at minimum 3 APs
Thereof, 425 devices were seen just once, while 370 were seen multiple times
301 devices were using MAC Randomization, 69 were not applying Randomization techniques