TL;DR
This paper introduces a novel access control scheme for Standalone Name Resolution (SNR) systems in Information-Centric Networking (ICN) to combat covert Content Pollution Attacks (CPA). By employing a decentralized voting mechanism and a malicious user blacklist, the system efficiently identifies and restricts attackers, significantly reducing the probability of successful collusion attacks.
What Problem Does This Paper Address?
Information-Centric Networking (ICN) relies on Name Resolution Systems (NRS) to map content identifiers to their locations. Standalone Name Resolution (SNR) is a popular approach for this, but it makes SNR nodes prime targets for attackers. One particularly insidious threat is the Content Pollution Attack (CPA). Unlike brute-force attacks, CPAs involve attackers slowly and subtly injecting invalid or undesirable content into the network. This consumes valuable SNR resources, increases resolution delays for legitimate users, and pollutes caches, all while being difficult to detect due to its low-speed, inconspicuous nature. Existing defenses often focus on cache security, neglecting the crucial role of the SNR system itself in preventing such attacks. This paper addresses the need for a robust mechanism within the SNR system to proactively identify and mitigate these covert attacks.
Methods
The core of the proposed solution is a decentralized revocation scheme based on a neighbor voting algorithm. When a user or an SNR node detects suspicious activity (e.g., repeated invalid content), they can initiate a revocation request. This request triggers a voting process among neighboring users.
Key components of the method include:
- Malicious User Blacklist (PKBL): A list maintained by SNR nodes to store revoked user public keys.
- Neighbor Voting Algorithm: Users with higher security weights and grades have more influence in the voting process. This incentivizes good behavior and participation in network security.
- Incentives and Punishments: Users are rewarded for reporting malicious behavior and penalized for contributing to attacks. This encourages active participation in maintaining network integrity.
- Public Key Grading: Users' public keys are assigned grades (Top, Professional, Personal, Basic) based on verification rigor and past contributions, influencing their voting weight.
- Security Weight: A dynamic value that increases with positive contributions (e.g., voting, reporting) and decreases with negative actions (e.g., publishing invalid content).
- Decentralized Revocation: The process is distributed among users, avoiding reliance on a central authority and improving efficiency.
- Collusion Resistance: The voting mechanism is designed to be robust against coordinated attacks where a group of users attempts to manipulate the voting outcome.
Datasets
Not discussed in the paper. The evaluation relies on simulations.
Key Insights and Ideas
- Shifting Defense Focus: The paper highlights that securing the SNR system itself, not just the caches, is critical for defending against CPA.
- Power of Decentralized Voting: Leveraging the collective intelligence of users through a voting mechanism can effectively identify and isolate malicious actors in a decentralized manner.
- Incentivizing Security Participation: By tying voting weight to user behavior and assigning grades to public keys, the system creates a strong incentive for users to act as responsible network citizens.
- Progressive Revocation: The scheme differentiates between minor infractions and malicious behavior, implementing a tiered revocation process that starts with content and escalates to user revocation.
- Robustness Against Collusion: The voting weight and threshold setting are crucial for mitigating collusion attacks, ensuring that a significant portion of the network must agree for a revocation to occur.
Key Results
The paper's experimental evaluation demonstrates the effectiveness of the proposed scheme:
- Reduced Collusion Attack Probability: With an attacker ratio of 0.5, the probability of a successful collusion attack is reduced to less than 0.1 when voting weights are continuously increased.
- Improved System Robustness: Even when half of the users in a domain are attackers, the system maintains a 90% probability of correct revocation.
- Efficient Revocation: The decentralized approach avoids the need for additional system components, making it more efficient.
- Communication Overhead: The communication overhead for a revocation event is analyzed and found to be manageable.
- Average Revocation Delay: The simulation shows that as the number of users increases, the average revocation delay decreases due to a higher density of active users.
Limitations and Future Work
- Simulation-Based Evaluation: While the simulations are extensive, the authors acknowledge the need for real-world implementation to gather more reliable data.
- Future Work: The authors plan to implement the scheme in an experimental ICN project (SEANet) to test its adaptability in actual network scenarios and further refine the scheme. They also aim to analyze the adaptability of the scheme for actual application scenarios.
Summary updated