Cyber criminals know that financial and e-commerce providers often use fraud detection systems that monitor for anomalous transactions and behaviors. In order to bypass these systems, fraudsters are continuously looking for ways to conceal behaviors that might be identified as potential fraud.

Evading Device Fingerprinting

The IBM Security research team recently came across a tutorial published in underground forums that explains how to circumvent fraud detection systems that use device identification to track and detect anomalous transaction and access patterns. This approach collects myriad session attributes to “fingerprint” the endpoint device, including IP address and type and version of browser and operating system. Using this information, fraud detection systems can, for example, detect when a single device is being used to place multiple orders with different user credentials — a practice typically indicative of fraud.

The tutorial recommends that fraudsters use commercially available virtual private network (VPN) or proxy services to hide the device’s IP address. It also offers guidelines on how to make sessions from a single computer appear as if they are originating from different computers with different browsers and operating systems.

According to the tutorial, users can achieve this with a few simple steps that involve downloading and installing an off-the-shelf utility for changing the browser’s user-agent header (where the operating system and browser information is contained). This technique prevents merchants from linking multiple transactions to a single device; each transaction appears to be coming from a legitimate customer who is using a unique computer and browser to place an order.

Finding Another Way

Because fraudsters can manipulate the browser’s user-agent header information easily, institutions should consider device fingerprinting solutions that rely solely on user-agent data unreliable. Device fingerprinting solutions should collect identification parameters via methods that are less likely to be manipulated by fraudsters. This stronger type of device fingerprinting solution can be a valuable component of a comprehensive, layered security solution.

The tutorial demonstrates that cyber criminals have achieved a sophisticated level of understanding of device fingerprinting techniques and are exploiting their knowledge to evade fraud-prevention systems that rely on the browser’s user-agent header to detect cyber crime. It is also a call to action for merchants and financial institutions that use device fingerprinting. They should make sure their solutions are collecting device information from a tamper-proof source.