Research Summary: SoK: Applying Blockchain Technology in Industrial Internet of Things

SoK: Applying Blockchain Technology in Industrial Internet of Things

Description:

Industrial Internet of Things (IIoT) platforms need to integrate blockchain technologies to overcome challenges such as poor interoperability, privacy, security, and resource constraints in addition to heterogeneous smart devices, networks, and data types.

TLDR

  • Traditional IIoT systems employ centralized cloud-based architectures that may be unsustainable for industries’ projected exponential growth.
  • The drawbacks of using a centralized architecture include having a single point of failure, massive data processing requirements, and lack user-access controls.
  • Decentralized architectures may address these shortcomings, however, integrating them requires addressing several challenges: standardization, interoperability, and scalability.
  • This study explores recent advances in tackling those shortcomings and examines several practical applications of blockchain architectures in IIoT platforms.
  • The author proposes three architectures for integration: IIoT-IIoT, IIoT-Blockchain, and Hybrid Approaches.

Core Research Question

What are the architectures that IIoT platforms need to integrate to sustain industrial advancement?

Citation

  • [1] G. Wang, “SoK: Applying Blockchain Technology in Industrial Internet of Things”, Cryptology ePrint Archive, 2021. [Online]. Available: https://eprint.iacr.org/2021/776.pdf. [Accessed Sept. 18, 2021]

Background

  • IIoT is a branch of the Internet of Things (IoT) that stems from the increasing prevalence of smart devices across industrial applications. At current rates of growth, there will be 75 billion IoT devices by 2025, with IIoT devices making up 17% of the total.
  • Since IIoT deals with massive amounts of data, the ubiquitous adoption of IoT devices will demand efficiency improvements. Current centralized solutions won’t be able to keep up with increasingly large and complex networks of IIoT devices.
  • Blockchain technologies may offset many of the traditional limitations of IIoT with decentralized mechanisms, immutability, and transparency.
  • Blockchain infrastructures can eliminate single points of failure and reduce overhead by distributing data processing [1].
  • Still, the scope of blockchain applications within the industrial sphere is relatively limited. Though integration presents considerable opportunities for advancement, no universal standards for integration currently exist.

Summary

Features and Challenges of IIoT

  • The primary function of IoT is connectivity. In industrial settings, IoT devices are Cyber-Physical Systems that link machines to control systems, interfacing physical operations with their cyber counterparts to reduce the need for human interference and allowing automation.
  • Most IIoT devices are lightweight nodes with resource constraints such as limited computing power, bandwidth, energy supply, memory, and storage.
  • IIoT devices often have unpredictable connections to one another due to device mobility, faulty wireless networks, and device standby settings.
  • Traditional IIoT platforms deploy centralized cloud-based infrastructures, which provide ample computing power and storage, yet when joined with IIoT technologies and devices they present significant drawbacks.
  • These include privacy and security vulnerabilities, access control issues, expenses and risks associated with third parties, and bottlenecking and scalability concerns.
  • IIoT platforms experience poor interoperability mainly due to their heterogeneous make-up. The uses of IIoT devices vary per application, as do network topologies and data types. Each industrial sector and organization may implement different protocols, making data and service sharing complex and inefficient.

Features, Benefits, and Challenges of Blockchain

  • Blockchain is a shared ledger that verifies and records transactions through decentralized technologies and distributed networks. All participating nodes within blockchain networks store identical copies of the ledger and maintain consistent timestamped records.
  • Data stored on the blockchain is immutable and tamper-resistant, as any alterations are reflected in successive blocks’ overturn. These qualities, coupled with blockchain’s transparency and traceability, make it a viable structure for business relations.
  • Blockchains generally apply asymmetric cryptography and digital signatures for security and privacy purposes.
  • Blockchain’s pseudo-anonymity provides a degree of user privacy, though privacy remains an open issue.
  • Blockchains are vulnerable to 51% attacks, partitioning attacks, and smart contract scripting exploits.
  • Scalability is an ongoing challenge. As the blockchain expands, performance diminishes and resource requirements increase.

Integration and Advantages of Blockchain and IIoT

  • As the number of IIoT devices on a network grows, the amount of data generated increases. Likewise, the demands on bandwidth and computational power also increase.
  • Consequently, centralized systems experience congestion and delay, resulting in expensive hardware solutions and steep maintenance overhead.
  • Using a blockchain’s decentralized structure as an overlay on existing IIoT frameworks can evenly distribute computational efforts among participants in a network.
  • Smart contracts stored on the blockchain serve as real-time auditors and may eliminate third-party actors in IIoT systems, thereby lowering expenses.
  • Blockchain’s composite layer acts as an interface that streamlines different data types and masks underlying heterogeneous layers. It connects the communication layer to industrial applications, bridging diverse mobile and industrial networks, enhancing interoperability, and facilitating reciprocal data exchange. Refer to Fig. 5.
  • Current industrial solutions employ lightweight IIoT devices to accelerate performance, yet this poses a growing risk to the industrial landscape. End-devices are subject to attacks since they cannot host robust security mechanisms and firmware updates are intermittent.
  • Integrating blockchain may enrich IIoT’s security systems through robust encryption primitives and multi-digital signature requirements. Moreover, blockchains are resistant to single points of failure attacks and provide a measure of fault tolerance.
  • Industrial applications typically employ consortium or private blockchains which enable smart contracts to regulate user access controls and data provenance. This improves privacy, security, and detection of unauthorized user access.
  • Omission of device authentication services potentially compromises IIoT operations.
  • Blockchain identification technologies can authenticate and monitor devices to assess and manage their integrity.

Challenges of Integration

  • IIoT devices have limited computing power, energy supply, and bandwidth capabilities. They cannot meet the requirements of blockchain mechanisms and sustain high levels of throughput.
  • The storage requirements of blockchain are too significant for lightweight nodes. However, devoid of the entire blockchain and its data, participating lightweight nodes cannot validate peer transactions.
  • Since industrial operations are time-sensitive, the latency introduced by consensus protocols may debilitate system performance and timestamping accuracy.

Method

The author employed a qualitative approach to explore architectures that IIoT platforms need to integrate to improve their current architectures. The researcher collected data through a comprehensive literature review consisting of 244 sources.

Results

While identifying integration challenges and potential solutions, the author provides some generalizations without necessarily connecting each problem to a solution and vice versa. Therefore, this summary only covers the issues and solutions that directly align.

Potential Solutions

  • The author proposes introducing structures similar to a blockchain that improve scalability and throughput and provide comparable services such as “decentralization and immutability” [1]. These include Directed Acyclic Graph, Tangle, and Greedy Heaviest-Observed Sub-Time.
  • On-chain storage may be reserved for critical data, whereas minute data may be stored off-chain and retrieved through distributed hash tables like Kademlia.
  • Latency can be improved by reducing transaction size and using powerful nodes such as edge devices to mitigate confirmation delays.

Discussion and Key Takeaways

Blockchain Storage and BaaS Platform

  • Blockchain communication requirements and massive data inhibit end-device performance and are too costly for on-chain storage solutions.
  • Decentralized blockchain storage run on cloud-based infrastructures can provide more efficient, secure, and cost-effective models for IIoT platforms as illustrated in Table IV.
  • Using Blockchain-as-a-Service (BaaS), industrial operations can integrate specific blockchain applications into existing cloud-based platforms with fewer complexities and investments. Essentially, this entails customizable blockchain capabilities to manage and store data with service providers that oversee blockchain operations. Table V. provides a list of various BaaS platforms.

Integration Approaches and Blockchain Selection

  • The author presents three models for blockchain integration: IIoT-IIoT, IIoT-Blockchain, and Hybrid Approaches. Table VII. provides a comparison of these approaches.
  • In the IIoT-IIoT model, blockchain has minimal interaction with IIoT and functions as an immutable archive, accounting for partial IIoT data and sustaining low latency levels, making it ideal for time-sensitive applications.
  • The IIoT-Blockchain model emphasizes the interconnectivity of blockchain and IIoT, recording all data via consensus mechanisms. This approach warrants more resources and values data as its greatest asset.
  • The Hybrid model secures partial data on the blockchain and requires a calculated design to capture critical data. Without heavy communication between IIoT and blockchain, it reduces burdens on constrained devices.

Implications and Follow-ups

Optimization on Performance

  • Industrial scenarios must sustain high levels of throughput for devices to function synchronously. However, existing solutions for blockchain integration increase throughput at the expense of scalability. Decreases in latency may also have adverse effects on scalability.
  • Most IIoT devices are resource-constrained, making direct integration impractical.
  • Integration solutions must accelerate scalability and throughput, lower latency, and accommodate the limitations of end-devices.

Scalability

  • Poor scalability interferes with the adoption of blockchain in IIoT networks.
  • Unstable wireless connections also contribute to scalability issues.
  • Areas of research that may surface potential solutions include sharding and side-chaining.

Security and Privacy

  • Blockchain and IIoT are predisposed to different vulnerabilities.
  • Smart contract scripting lacks standardization, making discrepancies and oversights exploitable.
  • IIoT wireless networks are susceptible to eavesdropping and Denial of Service attacks.
  • Privacy preservation is an open issue for blockchain-IIoT platforms. The author proposes “decentralized record-keeping that is completely obfuscated and anonymous by design” [1].

Editable Blockchain

  • Conventionally in a blockchain, all records should be stored by all nodes. However, some industrial data eventually becomes useless and can be deleted or transferred to secondary storage. Also, fraudulent data should be nullified.
  • Editability allows modification and deletion of irrelevant, incorrect, and generally undesirable data stored on the blockchain. However, there must be a compromise between editability and security of the system.

Edge Computing

  • IIoT devices frequently cannot meet the requirements posed by IIoT applications when blockchain is applied, specifically in the context of computation and networking. Their storage is limited, and there are also limits on their interoperability and authentication standards.
  • In edge computing, edge devices are placed close to edge servers. Edge servers are less powerful than cloud servers but have closer proximity to edge devices which in an ideal situation allows for minimization of latency and transmission delay.
  • Powerful gateways can also act as consensus nodes in the blockchain to resolve the storage and computing problems of lightweight devices.
  • This paper proposes a potential research direction in which blockchain technology is implemented on the IIoT edge and minimizes networking and computational overhead.

Standardization on Blockchain-Based IIoT

  • Lack of or inconsistency in standardization results in the inability to reach service agreements for integration processes.
  • The Institute of Electrical and Electronics Engineers (IEEE) and the International Standards Organization (ISO) have made attempted standardization efforts.
  • Blockchain standardization will help to redefine future technologies and assist both users and developers of blockchain.

Applicability

Industry 4.0

  • Total industrial automation can be realized through the convergence of IoT, Cyber-Physical Systems, and blockchain.
  • With their immutable services, decentralized systems, particularly blockchain, can enable secure, trustworthy, and cost-effective interactions among autonomous agents.
  • Quality of Service monitoring on the blockchain can improve latency through its updating requirements and coupled with smart contracts, it may actualize chaining in real-time.

Smart Manufacturing

  • Smart manufacturing predominantly relies on centralized infrastructures and third-party auditors, which pose considerable disadvantages: poor interoperability, increased overhead, and security vulnerabilities.
  • Interoperability may be improved through blockchain’s ability to meld disjointed IIoT systems into a distributed network.
  • Overhead can be lowered with smart contracts as real-time auditors.
  • Smart contracts can provide automated firmware updates to leverage security.
  • Through blockchain-enabled IIoT platforms, smart manufacturing could achieve advances like machine self-monitoring, self-diagnosis, and automatic maintenance requests as outlined in the Blockchain Platform for Industrial Internet of Things.

Smart Grid

  • Centralized Energy Management Systems (EMS) have been shown to be less efficient and secure for peer-to-peer (P2P) energy trading.
  • Decentralized blockchain EMS has potential for more efficient and trustworthy P2P large-scale energy trading, especially with regards to the security and privacy of energy exchange and transmission.
  • Decentralized EMS applying bi-level algorithms indicate more practical operations for renewable energy source distributed generators in microgrids.

Supply Chain

  • Blockchain may be implemented as a form of quality control in the supply chain. Its identification technologies translate physical assets to digital identities associated with immutable timestamps, providing traceability and tamper-resistant proofs.
  • A product can be traced from its source to the shelf to uphold food safety standards in the food industry.
6 Likes

Hi @rlj -

You mentioned that centralized systems experience congestion and delay, and blockchain’s decentralized structure is supposed to make it faster because it distributes computational to different participants.

How does this hold? Isn’t it commonly held that blockchain is computationally intensive and much slower than other data transmissions on the internet?

This is just speaking from my personal experience. Blockchain transactions take as long as half an hour to process. I would be concerned if any IoT device I own updates its data with that kind of latency, let alone the cost of computation.

Thanks!

2 Likes

@rlj thank you for posting this terrific addition to our scaling category. When we talked about this summary one-on-one you mentioned that IIoT + blockchain technologies had tremendous potential, to the point that the author describes it as being able to ignite a fourth industrial revolution. How come? How might these networked things improve industrial processes?

1 Like

@Twan I echo your concern about the latency built into consensus mechanisms. See also:

And:

However, with IIoT specifically, there seems to be a resistance to the “shift of mindset” necessary to flip constraints into opportunities. Note this passage:

This captures the point: Of course edge servers are less powerful than cloud servers. But when designed to have “closer proximity to edge devices” they can help conquer problematic latency.

This leaves me thinking that working with the benefits of decentralized, P2P structures is a “mindset” that conventional thinking hasn’t yet embraced.

2 Likes

Hi @Twan, thanks for your question. To clarify that matter, distributing computational efforts reduces expenses associated with “installation and maintenance in centralized infrastructures” and “networking equipment” [1]. That is not to imply it would reduce latency, though you raise a key issue in this research. Poor scalability largely interferes with blockchain-IIoT integration. Traditional IIoT platforms and blockchain both face this challenge. However, the latency introduced by blockchain’s decentralized mechanisms is unacceptable by industrial standards, which require massive amounts of data to be processed in a timely manner. As noted by the author, “solving scalability in blockchain will serve as a huge advance toward creating a practical decentralized infrastructure for IIoT applications” [1]. I would have to agree with the concerns you express. In its current state, blockchain comes at the expense of suboptimal industrial performance and with computational and networking overhead.

4 Likes

On the bright side, innovations in blockchain computation efficiency come out gradually. It’s not standing still. Researchers and engineers are working hard to reduce energy consumption. Ethereum even promised to reduce energy consumption by 99 percent on Ethereum 2.0.
If it’s going faster every day, maybe we can get lucky, and see blockchain overcome this barricade in the future.

2 Likes

Hi @rlj: Thanks for an excellent summary of a paper which reveals many of the contradictions presented by the history of IIoT, as well as the possibility of mitigating those contradictions with blockchain—if industry can bring itself to adopt the necessary new “mindset” around the issues at hand.

Here are a few key points that leapt out at me:

This highlights a key failing across most industries and OEMs: Attempting to achieve customer “lock-in” with core proprietary protocols and technologies rather than adopting common-sense standards and then differentiating products on top of those standards.

The assumption that all data is equally valuable, to be preserved forever, is simply in error. A huge percentage of industrial data can be collected and processed at the edge, acted upon appropriately, and then discarded.

The paper characterizes “edge” devices correctly…

…but it doesn’t say that if we adopted a fully networked, P2P, decentralized state of mind, the limited computing power, storage, etc. of IIoT devices might be seen as an opportunity not a constraint.

For example, the paper notes:

But then it goes on to say:

Why can’t these “robust encryption primitives” be implemented on edge devices themselves?

Finally, the paper makes an intriguing point:

This is a fascinating area. Can you say a little more about it?

1 Like

@GanouTeikun welcome to the forum! With a mechanical engineering background, I was thinking that this post may interest you.

3 Likes

@Gearlad how does this overlap with some of the work you and @Sean1992076 have been working on? @rlombreglia mentions embedding primitives in IoT devices, can you talk about some of the work your lab has done with turning IoT into lightweight ETH nodes and routers – would that be a solution ot some of the issues he’s brought up?

2 Likes

Having worked in this field, I had an experience that using blockchain for storing all the data is quite a terrible idea (from an implementational perspective). Instead signed Hash Tables can be the preferred way since we had like more than 50 GBs of data only from one industrial site per day. Imagine storing it as multiple copies stored on all participating nodes that too through consensus!

And for immutability of data (partial?), as long as at least one party can provide full copy of data along with the corresponding signed hash table from blockchain, we are good :slight_smile:

3 Likes