We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Employing blockchain and IPFS in WSNs for malicious node detection and efficient data storage.
- Authors
Saeed, Arooba; Javed, Muhammad Umar; Almogren, Ahmad; Javaid, Nadeem; Jamil, Mohsin
- Abstract
The presence of the malicious internet of things (IoT) devices in wireless sensor networks (WSNs), in the underlying work, is identified using a private blockchain and interplanetary file system (IPFS) based system. In the proposed model, the registration of IoT devices is being done via the IoT device manager. After registration, the IoT devices send a request to get services from the IoT device manager. At this stage, authentication is performed through a machine learning technique, known as logitboost (LB), which helps in identifying the IoT devices as malicious or legitimate. A verification report is forwarded to the IoT device manager upon detection. If the IoT device is found to be acting maliciously, it is penalized and vice versa. Thus, ensuring the IoT device manager provides the required service only to the legitimate IoT device. The data used in the proposed work is taken from the WSN dataset (WSN-DS), which is balanced via the synthetic minority oversampling technique. Additionally, the miner nodes' verification is performed via verifiable byzantine fault tolerance (VBFT). Also, IPFS is used to store the transaction data being shared between the IoT devices. The simulation results show that VBFT surpasses the proof of authority consensus mechanism by 15–20% in terms of transaction throughput. Furthermore, the LB classifier is found to surpass decision tree, nearest centroid, quadratic discriminant analysis and weighted average ensemble by 2%, 24%, 71%, 5% in terms of accuracy.
- Subjects
WIRELESS sensor networks; BLOCKCHAINS; BIOMETRIC identification; DISCRIMINANT analysis; FAULT tolerance (Engineering); DATA warehousing; DECISION trees
- Publication
Wireless Networks (10220038), 2024, Vol 30, Issue 4, p2313
- ISSN
1022-0038
- Publication type
Article
- DOI
10.1007/s11276-023-03648-3