Source-linked AI summary
Blockchain based trust & authentication for decentralized sensor networks
Axel Moinet, Benoît Darties, Jean-Luc Baril
TL;DR
Resource-constrained, decentralized WSNs lack a complete model for reliable peer authentication and trust management. The paper proposes BATM, a blockchain-based model storing authentication and trust information, and concludes that it can provide authentication and trust evaluation in self-organized, evolutive networks.
Problem
Existing models do not provide complete authentication and trust management for decentralized, resource-constrained WSNs.
Method
BATM uses blockchain storage for cryptographic authentication data and node trust information, with signed payloads supporting data verification.
Results
The paper shows a blockchain-based module and Human-like Knowledge based Trust model for authentication and trust evaluation in decentralized networks.
Takeaways & Limitations
The proposed model supports authentication and trust evaluation in self-organized and evolutive Wireless Sensor Networks.
Takeaways & Limitations
Malicious authenticated nodes can destabilize the system by overloading the network with valid payload submissions.
Abstract
from arXiv · showhide
Sensor networks and Wireless Sensor Networks (WSN) are key components for the development of the Internet of Things. These networks are subject of two kinds of constraints. Adaptability by the mean of mutability and evolutivity, and constrained node resources such as energy consumption, computational complexity or memory usage. In this context, none of the existing protocols and models allows reliable peer authentication and trust level management. In the field of virtual economic transactions, Bitcoin has proposed a new decentralized and evolutive way to model and acknowledge trust and data validity in a peer network by the mean of the blockchain. We propose a new security model and its protocol based on the blockchain technology to ensure validity and integrity of cryptographic authentication data and associate peer trust level, from the beginning to the end of the sensor network lifetime.
1 INTRODUCTION
The paper targets missing integrated authentication and trust management for resource-constrained, decentralized WSNs. It proposes a blockchain-based model for decentralized authentication and node trust information.
- WSNs support IoT development but face constraints in computation, hardware functionality, energy consumption, adaptability, and evolutivity.
- Existing research addresses network data security or node authentication and trust management, but not a complete model covering content access, security, privacy, and trust.
- The paper proposes blockchain storage for decentralized authentication and node trust information in an evolutive and adaptive model.
2 THE BLOCKCHAIN TECHNOLOGY
The paper presents blockchain as a hash-linked structure that preserves block continuity and supports generic decentralized storage beyond Bitcoin transactions. Structured payloads can replace transactions, but replication and bandwidth constrain storage design.
- Bitcoin models a blockchain as a linear sequence of blocks containing cryptographic hashes for the previous and current blocks.
- The block header includes the Merkle root of its transactions, linking that root into subsequent block construction.
- Changing one transaction changes the Merkle root and invalidates the block, while competing valid forks are resolved by selecting the longest blockchain.
- The blockchain can serve as generic decentralized storage by replacing transactions with arbitrary data payloads organized alongside block headers and constants.
- Changing one payload invalidates its block, but full-chain replication increases storage and bandwidth demands.
- The platform limits each block to 5 MB to preserve storage for security and trust information without overloading blockchain control traffic.
3 AUTHENTICATION AND TRUST IN WSN
The paper models decentralized WSNs as resource-constrained graphs whose nodes provide abilities and services. Its service-oriented model supplies the organizational basis for authentication and trust mechanisms.
- Network services model: The service model is adapted from Service Oriented Architecture and made compatible with CoAP for RESTful interaction with sensor networks.
- Network services model: A WSN is modeled as an undirected graph G = (V, E), with vertices representing nodes and edges representing transmission-range links.
- Network services model: The network model associates abilities and services with nodes to represent available resources.
- Network services model: A Network Node defines Node Properties and Node Abilities vectors.
- Network services model: Available Services define Abilities Dependencies, Resources Dependencies, and Resources Provider vectors.
- Network services model: Nodes store services in a Service Registry, and storage-capable nodes retain services they cannot deploy for future reuse.
4 BLOCKCHAIN AUTHENTICATION AND TRUST MODULE (BATM)
BATM uses blockchain storage for public keys, digital signatures, and peer information so network components can validate one another. Its design links authentication, block exchange, and trust management.
- Authentication foundations: PKI provides confidentiality through encryption, authentication through digital signatures, and peer-based identity validation through a web of trust.
- BATM design: BATM stores public keys, digital signatures, and peer information in a blockchain so network components can validate information about other nodes.
- BATM design: The BATM design covers authentication and block mining, block exchange rules, and trust management derived from their combination.
4.1 BATM authentication
BATM authenticates network nodes by storing and validating cryptographic credentials in a blockchain. Its block and payload rules control mining, credential updates, and payload integrity, while Blame and Ban payloads support trust management.
- Key management: BATM associates each network node and application service with cryptographic keys, using a master key to generate secondary encryption and signature keys.The paper identifies private-key management as critical because compromised keys can enable identity spoofing.
- Block mining: A block is valid only when it resolves the blockchain problem and contains a valid Miner Approval payload generated by an authenticated miner.The Miner Approval signs a random value from the previous block's approval payload and identifies the node that authorized the authentication entry.
- Credential authentication: At first registration, a node or application service broadcasts a Credential Payload containing the public keys required for network operation.Authentication is approved when an authenticated node includes that credential payload in a valid block.
- Credential authentication: Credential status can be updated through renew and revoke payloads, with replacement credentials included alongside revocations to preserve authentication continuity.A revoking node must provide a new credential payload, and miners try to include both in the same block.
- Payload validation: Every payload ends with an issuer-signed hash digest, allowing verification of payload data and distinguishing master-key signatures for revocations from subkey signatures for other payloads.This verification rule restricts payload validity to data authenticated by the appropriate issuer key.
- Trust management: BATM uses distinct payload types, including Blame and Ban payloads specifically for its trust management model.The blockchain therefore stores both authentication-related cryptographic material and information used to manage node trust.
4.2 BATM trust management
BATM uses the Human-like Knowledge based Trust model to maintain evolving node reputations from blockchain payloads, balancing mutual surveillance with centralized trust management. Reputation changes with event behavior over time and supports application-specific trust thresholds, while the system includes rules intended to improve stability against overload.
- HKT model: HKT maintains a reputation level for each node as a compromise between network-wide mutual surveillance and a trust center.The model is based on human-like behavior.
- Blockchain-based evaluation: Blockchain payloads indicate node behavior over time, supporting trust evaluation without a trust center and preventing data tampering or identity impersonation.The same principles apply to authentication servers, whose reputation is echoed across network nodes.
- Reputation evolution: HKT assigns reputation factors to payload events and decreases each event’s contribution over time using a continuous decreasing function such as e^-x.This makes reputation evolve as past events become less influential.
- Initial trust: Newly authenticated nodes receive a base trust value because they lack prior actions from which to compute reliable trust.The value is granted when a trusted node includes their credentials in the blockchain.
- Trust thresholds: Trust evaluation compares a node’s current reputation with minimum reputation thresholds for blockchain-related actions, weighted by the number of authenticated nodes and application restrictiveness.BATM defines thresholds including approval, blame, and ban levels.
- Simulation and stability: In the Figure 5 simulation, two blames trigger a node ban that becomes effective for 84 hours, although the simulation omits minimum trust levels required to fulfill actions.The simulation covers 500 hours with 3 NN.
- Simulation and stability: Malicious authenticated nodes can destabilize BATM by submitting many valid payloads, motivating specific payload-submission rules to improve system stability over time.The paper identifies overload by valid payloads as a system-level threat.
4.3 BATM payload rules
BATM limits payload activity and key lifetimes with timers and timeout rules, while storing application-specific values in the first blockchain block. The design also identifies an origin-block threat because an initial node can set mutable parameters that may destabilize the system.
- Timer rules: BATM uses three timers to regulate key renewals, blames, and the punishment period after banning a node.The timers are Trenew, Tblame, and Tbanrecover.
- Key timeout rules: Key-validity rules require Tsubkey to exceed Trenew but remain below 50×Trenew, while Tmasterkey must exceed 10×Tsubkey and not exceed 50×Tsubkey.These bounds address exposure to key-guessing attacks.
- Configuration: Timer values are application-dependent and stored in the blockchain’s first block for use by BATM.The first block stores mutable values used by the protocol.
- Origin block: The empty initial blockchain requires a special first block that any node can create from its own parameters.The system is intended to adapt after network events begin, and a malicious first node may be banned early.
- Origin block: A malicious initial node could insert values that make BATM misbehave because the first block contains all mutable parameters.The authors identify estimating BATM instability probability from these values as future work.
5 FUTURE WORK
Future work focuses on evaluating and extending BATM and HKT, including richer reputation factors, ability-aware trust decisions, and stronger resistance to denial-of-service attacks. The model is also considered for different network assumptions and a Linux implementation.
- Evaluation: The authors plan complete evaluation of each BATM component and the global model, followed by more adaptive algorithms using AS and NN characteristics.The current results are described as early reputation-evaluation results.
- Trust model: HKT may incorporate additional parameters, including node and service abilities, to improve trust and reputation evaluation.The authors also suggest weighting a blamer’s reputation and adding a time coefficient for network presence.
- Attack resilience: Requiring blames from different NNs or ASs is proposed as a way to increase resilience to denial-of-service attacks.
- Network models: A derivative model could use a network master key for blockchain supervision and to eliminate the first-block problem.The paper also considers vendor-specific networks using prior master-key signatures as membership proofs.
- Implementation: BATM may be included in the Multicast Services for Linux framework if simulation results meet expectations, but that framework remains under development without a release date.The planned implementation is intended as a real-world proof of concept for the overall design.
6 CONCLUSION
The paper presents BATM and HKT as a blockchain-based approach to authentication and trust management in autonomous wireless sensor networks. It describes blockchain storage for cryptographic keys and trust information, with implementation work still underway in MSL.
- Contribution: The paper proposes blockchain as secured decentralized storage for cryptographic keys and trust information in autonomous wireless sensor networks.BATM combines authentication mechanisms with trust evaluation in self-organized, evolutive networks.
- Implementation: The Service Oriented model is under development in the Multicast Services for Linux framework.