Abstract:Modern computers rely on USB and HDMI ports for connecting external peripherals and display devices. Despite their built-in security measures, these ports remain susceptible to passive power-based side-channel attacks. This paper presents a new class of attacks that exploit power consumption patterns at these ports to infer GPU activities. We develop a custom device that plugs into these ports and demonstrate that its high-resolution power measurements can drive successful inferences about GPU processes, such as neural network computations and video rendering. The ubiquitous presence of USB and HDMI ports allows for discreet placement of the device, and its non-interference with data channels ensures that no security alerts are triggered. Our findings underscore the need to reevaluate and strengthen the current generation of HDMI and USB port security defenses.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: Although the USB and HDMI ports of modern computers have built - in security measures, they are still vulnerable to threats based on passive power side - channel attacks. These ports can be exploited to infer the activities on the GPU, such as tasks like neural network computing and video rendering. By developing a custom - made device inserted into the USB and HDMI ports, the author shows that its high - resolution power measurement can successfully infer the GPU process. This attack method is covert and will not trigger any security alarms, emphasizing the need to re - evaluate and strengthen the current HDMI and USB port security defenses.
### Specific problem description
1. **Insufficiency of existing security measures**:
- Although the USB and HDMI ports have built - in security features (such as the USB disabling the auto - run function, the operating system notifying the user of the USB device's behavior of accessing the file system, and the HDMI port not supporting write operations), these measures are insufficient when facing passive attacks based on the power side - channel.
2. **Novelty of the attack**:
- Traditional attack methods usually directly measure the power consumption from the GPU or require specific system access rights. This paper proposes a new attack method, that is, inferring the activities on the GPU through the power consumption patterns of the USB and HDMI ports. This method does not require special permissions and does not interfere with the normal data channels, thus avoiding triggering security alarms.
3. **Potential risks**:
- This attack can obtain sensitive information about the operations on the GPU, such as the deep neural network architecture, video - rendering content, etc., by inserting a small and covert device into the computer's USB or HDMI port. Since USB and HDMI ports are common in modern computers, this attack method can be effectively implemented without installing software or additional hardware.
### Solutions and contributions
1. **GPU power side - channel inference**:
- A new method of inferring GPU activities through the power consumption patterns of the USB and HDMI ports is proposed. This method does not require directly measuring the power consumption of the GPU but utilizes the characteristics of peripheral devices being connected or sharing power.
2. **Design of power measurement prototype**:
- A power measurement tool based on the SparkFun ACS723 current sensor and Arduino board has been developed, which can collect high - resolution data at a speed of up to 350,000 samples per second for driving a machine - learning classifier to identify GPU activities.
3. **Evaluation of attack performance**:
- Through time - series data analysis and machine - learning models, the effectiveness of this attack method has been verified. Specifically, the attack can successfully identify the type of deep neural network running on the GPU and the video segments being rendered.
4. **Defense applications**:
- The potential defense applications of this technology in anomaly detection and malware identification are emphasized, for example, by detecting abnormal power consumption patterns to identify unauthorized resource - intensive activities.
### Summary
This paper reveals the security vulnerabilities of the USB and HDMI ports of modern computers when facing passive power side - channel attacks and proposes a novel attack method. By developing and testing a custom - made power measurement device, the author proves the feasibility and effectiveness of this attack, while emphasizing the importance of re - evaluating and strengthening the security defenses of these ports.