Blockchain technology with its decentralised, peer-to-peer architecture and immutable data records is often presented as a promising technology to remove centralised architectures of IoT (Internet of Things) systems, which would remove single points of failure and might increase users’ trust in IoT. However, the current technical and social challenges of realising IoT systems in combination with blockchains are hindering a widespread implementation. Various researchers have identified trust in IoT as a critical factor for success and acceptance of IoT across different sectors (Fernandez-Gago et al., 2017; Reyna et al., 2018).
Initially, blockchain technology was developed and designed for the specific purpose of executing transactions of digital currencies like Bitcoin and not for use with IoT systems (Nakamoto, 2008). It is essential to note in this context that both blockchain platforms and IoT systems are continually adapting to new requirements, which makes generalising from characteristics of both concepts difficult. The following focuses on the ‘Ethereum’ blockchain, as it has been identified as highly relevant for IoT applications due to the ability to execute smart contracts in IoT settings.
Executing transactions with the Ethereum blockchain incurs transaction costs that can be up to 300 times higher
The frequency of IoT transactions and the amount of communicated data between IoT devices and blockchains depend on the specific application. Autonomous vehicles might require a high volume of transactions and bandwidth, whereas IoT sensors in ‘smart homes’ only execute a few transactions per day (see Figure 1). Currently, Ethereum requires about 20 seconds to confirm transactions from an IoT device. This latency depends on the current traffic of transactions and is not acceptable in use cases where organisations and users expect an immediate confirmation, such as when buying or selling products and offering rapid services. Ethereum has announced an update to address the current challenge of scalability. The update aims at decreasing confirmation latency and increasing confirmation throughput to facilitate IoT scenarios with millions of devices.
Executing transactions with the Ethereum blockchain incurs transaction costs that can be up to 300 times higher than executing transactions with web services like Amazon SWF (Rimba et al., 2018). As a result, new technologies like ‘off-chain’ storage or different consensus mechanisms like Tangle by IOTA are emerging, which may lead to significantly lower transaction costs and thus, are a more attractive solution for IoT applications.
Another critical challenge of integrating IoT systems with blockchain is the heterogeneity of IoT devices’ technical capabilities, which range from low-performance IoT sensors to using full-nodes with IoT devices. Figure 1 depicts two potential IoT scenarios. The ‘smart home’ scenario involves mainly IoT sensors with low computational abilities that are not able to communicate directly with a blockchain due to restrained resources. To overcome this obstacle, the sensors communicate with a light node, which transfers captured data from IoT sensors to the blockchain. However, this use case is not fully able to profit from the benefits of encrypted communication of blockchains as the communication between IoT sensors and light nodes is often executed via less secure channels like ‘ZigBee’, which makes those IoT systems more susceptible to manipulations and cyber attacks.
The ‘smart home’ scenario can be used to illustrate a further challenge, which arises in IoT settings. IoT applications are collecting, transferring and analysing data to execute actions based on the deduced information, which highlights the importance of accurate and trustworthy input. Users and organisations might have to pay for or execute predefined actions depending on the collected data by IoT devices, which is the reason why trust in this data is fundamental for a broad application of IoT.
How can IoT participants be sure that an IoT sensor has measured accurate data or that data has not been manipulated before it was transferred to the light node and then immutably documented on the blockchain?
This challenge is described as the ’oracle problem’ (‘garbage in, garbage out’), as it is currently not possible to prove that accurate data was measured by IoT sensors. Abera et al. (2016) evaluated whether remote attestation of IoT devices can support the identification of compromised devices to ensure constant status monitoring and thus, to allow intervention when data collection is manipulated. Their proposed concept was valid for specific cases where the prover and specifier communicate directly, which may support the development of trust in IoT devices. For other cases involving indirect communication and heterogeneous devices, viable solutions to this problem must be developed, which further delays the emergence of trust in IoT.
The integration of IoT into blockchain platforms can be described as a continuous adoption and learning process, as new platforms with different blockchain characteristics are constantly evolving to cater to the specific needs of IoT.
In addition to continuous learning, researchers are developing frameworks to understand the emergence of trust between users and IoT devices, since only solving the technical problems of combining both technologies will not necessarily lead to wider adoption (Fernandez-Gago et al., 2017). Human and social factors have a significant influence on how technology is used and gets appropriated by users. Previous experience regarding the behaviour of IoT devices in the past helps humans to gain an understanding of how these devices are acting in different contexts and can be seen as a prerequisite for developing user trust. If no prior experience of specific devices is available, the development of initial trust can be more challenging, as it requires users, which are open-minded and venturesome.
In his TEDTalk, Daniel Price illustrates that IoT devices and their users create new social landscapes with social relationships, since IoT devices will eventually learn about user behaviour and will tailor their services to the user's specific needs. Users will only be willing to accept those IoT devices in their private lives if they are confident that the devices are transparent, reliable and trustworthy.
However, the challenge of building trust is not solely about relationships between humans and IoT devices. Many IoT devices interact with each other in an IoT system and must be able to assess whether other devices are trustworthy. Predefined evaluation mechanisms consider various technical parameters and execute an automated trust estimation to determine whether IoT devices should mutually exchange sensitive data in transactions (Nguyen et al., 2017; Abera et al., 2016).
To conclude, the current discussion emphasises the various challenges that need to be addressed in the technical and social realms before a wide-ranging application and use of IoT in combination with blockchain will be feasible. The promise of removing central authorities from IoT systems and enabling more autonomous services with smart contracts may significantly change products, services and business models in the future. However, this can only be achieved once the technical challenges of limited scalability, high transaction costs and incompatibility of low-performance IoT devices with most consensus mechanisms are solved. Furthermore, future research should aim to overcome the social challenge of establishing user trust in IoT, as a fundamental condition for broad adoption and acceptance among the users.