Real-time Cyberattack Detection with Collaborative Learning for Blockchain Networks

TL;DR

This work tackles cybersecurity in blockchain networks by proposing a real-time collaborative detection approach that preserves data privacy across distributed mining nodes. A private Ethereum-based BNAT dataset is created to train and evaluate a Deep Belief Network (GRBM/RBM) classifier, with gradient-sharing enabling decentralized global learning. Results show near-centralized performance, achieving up to 97% detection accuracy in simulations and around 95% real-time accuracy, while avoiding raw data sharing. The approach enhances security for decentralized blockchain deployments by reducing overhead and preserving privacy, and can be extended to more attack types and learning strategies.

Abstract

With the ever-increasing popularity of blockchain applications, securing blockchain networks plays a critical role in these cyber systems. In this paper, we first study cyberattacks (e.g., flooding of transactions, brute pass) in blockchain networks and then propose an efficient collaborative cyberattack detection model to protect blockchain networks. Specifically, we deploy a blockchain network in our laboratory to build a new dataset including both normal and attack traffic data. The main aim of this dataset is to generate actual attack data from different nodes in the blockchain network that can be used to train and test blockchain attack detection models. We then propose a real-time collaborative learning model that enables nodes in the network to share learning knowledge without disclosing their private data, thereby significantly enhancing system performance for the whole network. The extensive simulation and real-time experimental results show that our proposed detection model can detect attacks in the blockchain network with an accuracy of up to 97%.

Figures (5)

• Figure 1: Our proposed collaborative cyberattack detection model. The detection modules are first trained by their local data. They are then used to detect attacks for incoming traffic of blockchain networks before putting them into the mining nodes.
• Figure 2: The architecture of a DBN. This architecture includes multiple GRBM and RBM layers for classifying blockchain network traffic.
• Figure 3: Experiment setup in our laboratory. This experiment includes three Ethereum nodes and three servers in a network.
• Figure 4: Visualization using PCA for collected datasets.
• Figure 5: The convergence of learning models in three schemes.