Technology

pioneering new forms
of authentication for wireless devices

pioneering new forms of authentication for wireless devices

LEVL’S AI-driven approach provides software solutions for device profiling and identification

Over the last 10 years the number of wireless devices we rely on at home and in businesses has skyrocketed. In many cases, especially in IoT, security has been more of an afterthought than a key design requirement.

It has become clear across industries that these vulnerabilities need to be closed as hackers have targeted them. Today, many of these devices can be hacked with a quick internet search and a $99 purchase on Amazon. One of the unique aspects of IoT is that it combines the physical space with the cyber space. When these two spaces merge, new security challenges arise as attackers can exploit the physical space to overcome the cyber space and gain access to valuable assets. The major threat is to communication types where close, physical proximity is relied on as part of overall device security. Essentially, scenarios where devices share the same physical space and communicate with each other directly over the air. These types of communication can be susceptible to “Relay” and “Replay” attacks.

The industry’s only solution that can holistically solve “unsolvable” sensor attacks, such as:

  • Evil twin attacks
  • Attacks on physical access management readers
    and perimeter security
  • Attacks on smart locks
  • Attacks on keyless car entry systems
  • Credential theft
  • GPS spoofing

how we do it

Unique RF fingerprinting technology

Every wireless device has some small variations and anomalies that can be captured when the device transmits wireless signals. A “fingerprint” is a set of these signal anomalies which uniquely defines a device’s physical identity. These parameters can be passively extracted, for example, by analyzing the signals transmitted by the wireless device. LEVL’s technology can learn the unique “fingerprint” of every device once, and then use that fingerprint to authenticate that device’s identity.

Signal waveform analysis

Unlike solutions that process the digital data passed in the network, our technology does not require access to the digital data transmitted. Instead, it only analyzes physical layer information, such as the signal waveform used to transmit the data over the radio-frequency spectrum. The fact that no access to the data is required means that there are no privacy or network performance concerns when deploying our technology.

Software only solution for added security and functionality

Our software is embedded into the existing sensor/reader’s firmware/OS to upgrade the security and add extra functionality. Unlike many existing device authentication solutions in the market which incorporate proprietary hardware, we provide a software only solution which makes the integration faster and the adoption frictionless.

No device modification required – “Agentless” clients

LEVL’s standard software architecture includes two parts:

•  A cloud backend

  A software that runs locally, and only on the receiver side 

Our software solutions enable the hosting receiver to map, identify, authenticate, and track wireless devices in its proximity. One of the most powerful properties of our fingerprinting technology is that it does not require any hardware or firmware modification on the device that is being authenticated and is purely passive and agent-less. The entire intelligence of our solutions runs on the receiving side only. This makes our solution highly scalable.

Rapid Machine Learning Approach

LEVL’s technology features a machine learning approach, using statistical modeling to accelerate a system’s learning process. Each use-case has its own requirements: for example, for the specific use-case of BLE fingerprinting of mobile devices, our system requires only two minutes to build a unique physical ID for the device and a couple of seconds to authenticate its identity in the runtime.

levl 2020 © all rights reserved

made by netrise