Source-linked AI summary
Blockchain in internet of things: Challenges and Solutions
Ali Dorri, Salil S. Kanhere, Raja Jurdak
TL;DR
IoT needs security and privacy mechanisms that accommodate decentralized networks and resource-constrained devices, while conventional blockchain introduces computational, bandwidth, and delay overheads. The paper proposes a hierarchical blockchain-based architecture spanning smart homes, an overlay network, and cloud storage, with distributed trust mechanisms. Its qualitative evaluation indicates that the architecture addresses most security and privacy threats with constant best-case overhead and worst-case transaction scaling by clusters rather than nodes, while further attack-resistance questions remain.
Problem
IoT security and privacy are difficult to address because its decentralized topology and resource-constrained devices limit conventional approaches, while blockchain can impose substantial computational, bandwidth, and delay overheads.
Method
The paper proposes a hierarchical architecture using local, overlay, and cloud blockchains, coordinated through access policies, multisignature transactions, encryption, and distributed trust methods.
Results
The qualitative evaluation reports that the architecture handles most security and privacy threats, with constant best-case overhead and worst-case transaction scaling by clusters rather than nodes.
Takeaways & Limitations
The architecture provides a lightweight, secure, and private framework broadly applicable to multi-tiered IoT network topologies while retaining most blockchain benefits.
Takeaways & Limitations
Further work is needed to reduce vulnerability to denial-of-service, modification, and 51% attacks in establishing distributed trust.
Abstract
from arXiv · showhide
The Internet of Things IoT is experiencing exponential growth in research and industry, but it still suffers from privacy and security vulnerabilities. Conventional security and privacy approaches tend to be inapplicable for IoT, mainly due to its decentralized topology and the resource-constraints of the majority of its devices. BlockChain BC that underpin the crypto-currency Bitcoin have been recently used to provide security and privacy in peer-to-peer networks with similar topologies to IoT. However, BCs are computationally expensive and involve high bandwidth overhead and delays, which are not suitable for IoT devices. This position paper proposes a new secure, private, and lightweight architecture for IoT, based on BC technology that eliminates the overhead of BC while maintaining most of its security and privacy benefits. The described method is investigated on a smart home application as a representative case study for broader IoT applications. The proposed architecture is hierarchical, and consists of smart homes, an overlay network and cloud storages coordinating data transactions with BC to provide privacy and security. Our design uses different types of BCs depending on where in the network hierarchy a transaction occurs, and uses distributed trust methods to ensure a decentralized topology. Qualitative evaluation of the architecture under common threat models highlights its effectiveness in providing security and privacy for IoT applications.
Introduction
IoT enables fine-grained data collection and personalized services but creates serious security and privacy concerns. Blockchain offers relevant decentralized security properties, yet its computational, latency, scalability, and bandwidth costs motivate a lightweight IoT architecture.
- Motivation: IoT connects embedded and cyberphysical systems to collect highly granular information about people and environments.This data can improve efficiency and support services such as pervasive healthcare and smart cities.
- Motivation: Dense collection and dissemination of IoT data can reveal private behavior and lifestyle patterns.
- Blockchain Opportunity: Blockchain provides decentralization, anonymity, and security that suit IoT’s distributed and heterogeneous networks.These properties can reduce single points of failure, preserve user privacy, and secure networks involving untrusted parties.
- Challenges: Blockchain adoption in IoT is challenged by intensive mining, long block-generation times, poor scaling, and significant traffic overhead.These costs conflict with resource restrictions, low-latency requirements, large node populations, and bandwidth limits.
- Contribution: The paper proposes a lightweight, decentralized blockchain-based IoT architecture and illustrates it with a smart-home example while targeting diverse IoT use cases.The architecture is presented as application-agnostic and broadly suited to different applications.
Block-based IoT Architecture
The proposed architecture organizes IoT operation around smart homes, with local blockchains managed by resource-capable devices. These blockchains chain device transactions and enforce owner-controlled access policies without proof-of-work mining.
- Architecture: The architecture includes smart homes, an overlay network, and cloud storage, supporting remote access and device data storage.The smart-home setting includes devices such as thermostats, bulbs, cameras, and sensors.
- Local Blockchain: A local blockchain is mined and stored by one or more resource-capable devices that remain online.A smart hub or home computer can serve this role.
- Local Blockchain: The local blockchain chains transactions for each device and uses an owner-controlled policy header as an access-control list.Devices communicate only when the owner permits access through shared keys.
- Local Blockchain: Local blocks are appended without proof-of-work or other puzzles to reduce blockchain overhead.Transactions are still grouped and mined in blocks, with each block pointing to its predecessor and carrying forward policy information.
Overlay Network
The overlay network links resource-capable nodes through anonymous connections and clusters them to reduce communication overhead and delay. Cluster heads maintain access-related lists and an overlay blockchain for relevant transactions.
- Network Organization: Overlay nodes include smart-home miners, other high-resource home devices, smartphones, and personal computers.Nodes use Tor for additional IP-layer anonymity, and users may operate multiple overlay nodes.
- Network Organization: Overlay nodes form clusters, each electing a Cluster Head, and may change clusters when delays become excessive.Clustering is intended to decrease network overhead and delay.
- Cluster Control: Cluster heads maintain requester and requestee public-key lists plus a forward list for transactions sent to other cluster heads.These lists help determine whether access transactions remain within a cluster or are forwarded.
- Overlay Blockchain: An overlay blockchain maintained by all cluster heads records multisignature cloud-storage and access transactions.Cluster heads independently decide whether to retain or discard received blocks based on communication about the transaction.
Cloud Storage
Cloud storage organizes user data into numbered blocks and authenticates access with the block number and data hash. Shared-key encryption hides block numbers, while hash chaining supports exclusive access and continued ledger updates.
- Data Storage: Cloud storage groups user data into identical blocks associated with unique block numbers and stores each block with the data hash.Received packets are stored in first-in-first-out order.
- Authentication: Users authenticate to cloud storage by supplying the block number and hash of the stored data.Successful retrieval using both values authenticates the user.
- Privacy: The new block number is encrypted with a shared key derived from generalized Diffie-Hellman.The design relies on the key holder being the only party who knows the block number.
- Integrity: Collision-resistant hashes and secret block numbers are used to restrict access and chain fresh data to an existing ledger.
Transaction Handling
The architecture handles transactions through defined processes for storing, accessing, and monitoring IoT data.
- Transaction handling is introduced as the process governing how data transactions are handled.
Storing
Devices store data in local, shared, or cloud storage according to a defined policy, with cloud storage supporting smart services in the thermostat example.
- Devices may store data in local, shared, or cloud storage based on the defined policy.
- A smart thermostat typically stores data in cloud storage so the service provider can implement smart services.
- During bootstrapping, cloud storage returns a pointer to the first data block.
Accessing
Service providers access stored device data through signed multisig transactions, while miners tailor responses to authorization and query needs.
- A service provider creates and signs a multisig transaction to request stored data for a time period or an entire device chain.
- The transaction requires signatures from the requester and the smart home’s miner before being sent to the service provider’s chain hub.
- Authorized users or service providers receive block numbers and data hashes, while other requesters receive the minimum data needed to answer their query.The system can use techniques such as adding noise or a safe answer for restricted responses.
- Forwarding chain hubs, including the requester’s and requestee’s hubs, record the transaction in their blockchains.
Monitoring
Monitor transactions let a smart home owner obtain real-time information from a device, including continuously delivered streams such as live camera data.
- A monitor transaction lets the miner request real-time data from the requested device and send it to the requester.
- The mechanism supports checking a smart thermostat’s current configuration.
- Data may be sent continuously, as when a user views a live camera stream.
- Figure 2 identifies separate processes for store, access, and monitor transactions.
Distributed Trust
The architecture distributes trust across a hierarchical IoT network using distinct blockchain roles and transaction-handling methods. Its security mechanisms address common attacks while its overhead generally scales with clusters rather than nodes.
- Trust mechanism: Trust ratings among cluster heads use direct and indirect evidence under the Beta Reputation System.Direct evidence comes from verifying a block mined by another cluster head; receiving that block through another cluster head provides indirect evidence.
- Tiered blockchain design: The architecture employs local, shared, and overlay blockchains across different tiers instead of using Bitcoin’s blockchain uniformly.Table 1 summarizes differences between Bitcoin’s blockchain and the blockchains used in the proposed architecture’s tiers.
- Threat handling: Requester and requestee public-key lists in cluster heads reduce the effect of denial-of-service attacks by filtering or forwarding transactions.Cluster heads can also block a public key after repeated unsuccessful access requests, although attacks using different public keys may succeed.
- Threat handling: Users can change transaction identifiers and public keys for each transaction, helping prevent linking transactions to a real-world identity.The architecture also permits users to send arbitrary transactions to the overlay network.
- Threat handling: Chaining each storage block to a further block prevents a requester from attaching its data to the user’s data.The miner creates the additional chain link before providing the requester with the block-number and hash.
- Overhead: Except for new miner joining, actions scale at worst with the number of clusters, which grows more slowly than the number of network nodes.For same-cluster access transactions, computation and memory overhead are constant, while packet overhead and delay depend on hops between the smart home and storage.
Conclusion
The proposed blockchain-based IoT architecture addresses security and privacy while accounting for constrained devices. Its qualitative analysis reports favorable overhead scaling, but several attack vulnerabilities remain open research questions.
- Conclusion: The architecture handles most security and privacy threats while considering the resource constraints of many IoT devices.The paper frames IoT security and privacy as critical to realizing the technology’s expected societal and economic benefits.
- Conclusion: At best, the architecture has constant performance overhead; at worst, most transactions scale with the number of clusters rather than nodes.The architecture is presented in a smart-home context but is described as broadly applicable to multi-tiered IoT network topologies.
- Open questions: Further work is needed to reduce vulnerability to denial-of-service, modification, and 51% attacks in establishing distributed trust.The paper identifies IoT’s broadcast medium, decentralization, and resource constraints as challenges for addressing these questions.