💻 computer science

Multi-sensor Privacy Data Tampering Prevention Method under Flood Attacks

This paper proposes a multi-sensor privacy data tampering prevention method that combines distributed authentication, an improved AES-encrypted T-box, and differential variance-based flood attack detection to ensure the security, integrity, and confidentiality of data during transmission and storage.

Original authors: Aiping Liu, Zongbo Wu, Junjie Wu, Tianjun Ma

Published 2026-06-29

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Aiping Liu, Zongbo Wu, Junjie Wu, Tianjun Ma

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a city full of tiny, helpful messengers (sensors) that constantly send important notes about the weather, traffic, or your health to a central post office. These notes often contain private secrets. However, a "Flood Attack" is like a villain sending millions of fake, junk letters to the post office at once. This clogs the roads, exhausts the mail carriers' energy, and creates chaos where the bad guys can sneak in, steal the real letters, or even swap the private notes with fake ones without anyone noticing.

The authors of this paper, Aiping Liu and her team, propose a three-step "super-security system" to stop this from happening. Here is how their method works, explained simply:

1. The "Secret Tunnel" and the "Double-Check" (Distributed Authentication & Covert Channels)

First, the system sets up a Distributed Authentication Architecture. Think of this as having a team of security guards at every single house (sensor) instead of just one guard at the main gate. Before a letter leaves a house, the guards check its ID and make sure it's really from that house and hasn't been altered.

But what if the main roads are clogged with junk mail? The authors build a Covert Transmission Channel. Imagine this as a secret tunnel hidden inside the walls of the city.

How it works: Instead of putting the private letter in a visible envelope on the street, the system hides the private data inside the "timing" of the delivery. It's like sending a secret message by tapping a specific rhythm on a door, while the actual delivery truck looks like it's just carrying normal trash. To an outsider (the attacker), it just looks like normal traffic, but the receiver knows exactly how to decode the secret rhythm to get the real message. This makes it incredibly hard for the attacker to find or tamper with the data.

2. The "Unbreakable Scramble" (Improved AES Encryption)

Even if the attacker manages to peek into the secret tunnel, they shouldn't be able to read the message. To prevent this, the team uses an Improved AES Algorithm with a special "T-box."

The Analogy: Think of your private data as a clear glass of water. The attacker could easily see what's inside. The encryption algorithm takes that glass and puts it through a high-speed blender (the T-box) that scrambles the water into a complex, unrecognizable mist.
The Process: The data is chopped up into tiny 64-bit pieces, scrambled using a unique key (like a secret recipe), and then reassembled. Only the person with the exact matching key can un-blend the mist back into the original water. This ensures that even if the data is stolen, it looks like random noise to the thief.

3. The "Smart Traffic Cop" (Differential Variance Change Rate)

Finally, the system needs to stop the flood of junk mail before it overwhelms the whole city. The authors created a Dynamic Traffic Monitoring System that acts like a super-smart traffic cop.

How it works: This cop doesn't just count cars; it watches how fast the traffic changes. If the flow of cars suddenly spikes in a weird, unnatural way (like a sudden surge of 10,000 cars in one second), the cop knows it's a flood attack, not just a busy day.
The Action: Using a mathematical formula called "Differential Variance Change Rate," the system calculates if the traffic pattern is suspicious. If it is, the traffic cop instantly flips a switch (using a controller API) to close the specific gate where the junk mail is coming from, blocking the attack while letting the real secret messages pass through.

The Results: Did it work?

The researchers tested their system against four types of "flood attacks" (like sending millions of fake requests).

Old Way: When using normal roads, the system lost a lot of data and was easily tampered with, especially during heavy attacks.
New Way: With their secret tunnels, scrambled messages, and smart traffic cops, the system kept the data safe. The "security score" stayed very high (close to 100%), and very few messages were lost. The scrambled data looked so chaotic that it was nearly impossible for attackers to guess what was inside.

In short: This paper proposes a way to protect private sensor data by hiding it in secret tunnels, scrambling it so it's unreadable, and using a smart system to instantly block the massive waves of junk traffic that try to break the system.

Technical Summary: Multi-sensor Privacy Data Tampering Prevention Method under Flood Attacks

1. Problem Statement

The rapid expansion of multi-sensor systems in fields such as intelligent transportation, environmental monitoring, and healthcare has led to the collection of massive amounts of data, including sensitive personal privacy and critical infrastructure status. However, sensor nodes are resource-constrained and rely on distributed wireless communication, making them vulnerable to network attacks. Specifically, flood attacks (Denial of Service) overwhelm target nodes with excessive data packets, causing resource exhaustion and communication paralysis. This environment creates a high risk for data tampering during transmission and storage, as real-time detection of malicious behavior becomes difficult. Existing solutions, such as localized differential privacy, blockchain-based transaction optimization, and privacy computing, face limitations including unreasonable privacy budget allocation, new security risks from secret sharing, metadata exposure, and excessive computational overhead.

2. Methodology

The proposed method addresses these challenges through a three-stage framework designed to ensure the security, integrity, and confidentiality of multi-sensor privacy data under flood attack conditions.

A. Distributed Authentication and Covert Transmission

To prevent tampering at the source and during transmission, the authors employ a Distributed Authentication Architecture:

Stream Authentication: Multi-sensor data streams are preliminarily authenticated using specific authentication parameters ( $Q_i$ ) to verify source and integrity.
Security Evaluation: A differential calculation of data stream features is performed to evaluate security levels. If the feature variance indicates low security, an enhanced data function is applied to boost the data's anti-tampering capability.
Covert Transmission Channel: To avoid embedding private data in standard protocol fields (which is insecure), a time-based hidden transmission channel is constructed. Private data is pre-processed, encoded, and embedded into ordinary node source data using a time-characteristic embedding algorithm. This creates a covert channel where data is transmitted as feature packets, significantly reducing the visibility of the private data to attackers.

B. Encryption via Improved AES Algorithm

To further strengthen confidentiality within the covert channel, an Improved AES algorithm is utilized to design an encrypted T-box:

Process: The algorithm operates in two rounds: data row obfuscation and column obfuscation.
Implementation: Private data is segmented into 64-bit blocks. An initial key is generated and expanded. The plaintext state matrix undergoes obfuscation (row and column) followed by encryption using the expanded key set.
Decryption: At the receiving end, identity authentication is performed using the original and expanded keys. Upon successful authentication, the T-box decrypts the ciphertext back to plaintext.

C. Flood Attack Detection and Defense

To prevent resource exhaustion that could invalidate the above protections, a dynamic traffic monitoring mechanism based on the Differential Variance Change Rate is implemented within an SDN (Software-Defined Networking) architecture:

Traffic Monitoring: The controller collects real-time traffic data ( $V(t)$ ) from multi-sensor devices.
Anomaly Detection: The system calculates the differential variance ( $D_{var}(t)$ ) between current and previous traffic cycles to capture subtle, rapid changes indicative of flood attacks.
Decision Logic: A membership function $I(t)$ determines if traffic is abnormal. To avoid false positives, the system tracks the continuous duration of anomalies ( $s$ ) against a tolerable limit ( $d$ ).
Defense: If an attack is confirmed (i.e., $I(t)$ remains above zero for $d$ cycles), the controller uses its API to issue flow table entries that block the specific attack traffic, ensuring the continuity of legitimate data transmission.

3. Key Contributions

Integrated Defense Architecture: The paper proposes a holistic method combining distributed authentication, covert channel construction, and improved encryption, specifically tailored for the constraints of multi-sensor networks under flood attacks.
Time-Based Covert Channel: The construction of a hidden transmission channel based on time characteristics allows for the concealment of private data within ordinary traffic, enhancing resistance to interception and tampering.
Improved AES T-box: The design of an encryption T-box using an improved AES algorithm provides a two-round obfuscation process (row and column) to ensure high-confusion ciphertext.
Dynamic Traffic Defense: The introduction of the differential variance change rate algorithm allows for fine-grained, real-time detection of flood attacks, enabling the SDN controller to dynamically block malicious flows before they exhaust system resources.

4. Experimental Results

The method was evaluated in a testbed involving temperature, humidity, pressure, and light sensors under various flood attack scenarios (TCP SYN, UDP, ICMP, and Mixed floods).

Anti-Attack Capability: Compared to traditional transmission channels, the proposed covert transmission channel demonstrated significantly higher anti-attack indices (consistently near 100%) across all attack types, whereas traditional channels fluctuated widely and dropped as low as 40% in mixed flood scenarios.
Security Index and Packet Loss: The proposed method achieved a network security index close to 1 (indicating high integrity) and maintained a packet loss rate around 5%. This performance was superior to methods based on privacy protection alone, blockchain/privacy computing fusion, and improved blockchain, which showed lower security indices and higher fluctuation.
Information Entropy: In terms of ciphertext confusion, the proposed method's information entropy values closely matched the "most ideal" information entropy curves. In contrast, the blockchain-based methods showed significant deviations and lower entropy, indicating weaker encryption states in complex scenarios.

5. Significance and Claims

The paper claims that the proposed method effectively ensures the security, integrity, and confidentiality of multi-sensor privacy data during both transmission and storage, even in the presence of flood attacks. By integrating distributed authentication, covert transmission, and dynamic traffic defense, the method overcomes the limitations of existing single-layer defense schemes. The authors assert that this approach provides a stable and reliable solution for multi-sensor systems operating in complex network environments, preventing data tampering and ensuring the continuous operation of critical infrastructure. The work is presented as a comprehensive solution that addresses the specific vulnerabilities of resource-constrained sensor nodes against denial-of-service and tampering threats.