A Survey of IoT Applications in Blockchain Systems:

Architecture, Consensus and Traffic Modeling

LAPHOU LAO, The Hong Kong Polytechnic University, China

ZECHENG LI, The Hong Kong Polytechnic University, China
SONGLIN HOU, The Hong Kong Polytechnic University, China
BIN XIAO∗ , The Hong Kong Polytechnic University, China
SONGTAO GUO, Chongqing University, China
YUANYUAN YANG, Stony Brook University, America
Blockchain technology can be extensively applied in diverse services, ranging from online micro-payment,
supply chain tracking, digital forensics, health-care record sharing to insurance payment. Extending the
technology to the Internet of things (IoT), we can obtain a verifiable and traceable IoT network. Emerging
research in IoT applications exploits blockchain technology to record transaction data, optimize current
system performance or construct next-generation systems, which can provide additional security, automatic
transaction management, decentralized platforms, and offline-to-online data verification, etc. In this paper,
we conduct a systematic survey of the key components of IoT blockchain and examine a number of popular
blockchain applications.
In particular, we first give an architecture overview of popular IoT-blockchain systems by analyzing their
network structures and protocols. Then, we discuss variant consensus protocols for IoT blockchains, and
make comparisons among different consensus algorithms. Finally, we analyze the traffic model for P2P and
blockchain systems and provide several metrics. We also provide a suitable traffic model for IoT-blockchain
systems to illustrate network traffic distribution.

CCS Concepts: • General and reference → Surveys and overviews; • Networks → Network architec-
tures; Network protocols; • Theory of computation → Models of computation.

Additional Key Words and Phrases: Blockchain, IoT, architecture, consensus, traffic modeling

ACM Reference Format:

Laphou Lao, Zecheng Li, Songlin Hou, Bin Xiao, Songtao Guo, and Yuanyuan Yang. 2019. A Survey of IoT
Applications in Blockchain Systems: Architecture, Consensus and Traffic Modeling. 1, 1 (November 2019),
32 pages. https://doi.org/10.1145/nnnnnnn.nnnnnnn

∗ Corresponding Author

Authors’ addresses: Laphou Lao, Department of Computing, The Hong Kong Polytechnic University, Hong Kong, China,
[email protected]; Zecheng Li, Department of Computing, The Hong Kong Polytechnic University, Hong Kong,
China, [email protected]; Songlin Hou, Department of Computing, The Hong Kong Polytechnic University, Hong
Kong, China, [email protected]; Bin Xiao, Department of Computing, The Hong Kong Polytechnic University,
Hong Kong, China, [email protected]; Songtao Guo, Department of Electronic and Information Engineering,
Chongqing University, China, [email protected]; Yuanyuan Yang, Department of Computer Science, Stony Brook
University, New York, America, [email protected].

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee
provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and
the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored.
Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires
prior specific permission and/or a fee. Request permissions from [email protected].
© 2019 Association for Computing Machinery.
XXXX-XXXX/2019/11-ART $15.00
https://doi.org/10.1145/nnnnnnn.nnnnnnn

, Vol. 1, No. 1, Article . Publication date: November 2019.

2 L.Lao et al.

1 INTRODUCTION
Blockchain technology has the potential to become ubiquitous in financial services and industries. As
the core mechanism of publicized cryptocurrencies, such as Bitcoin [Nakamoto 2008] and Ethereum
[Wood 2014], blockchain has demonstrated its intrinsic characteristics, including decentralization,
anonymity, and auditability. A large number of blockchain applications has been deployed. The
worldwide estimated expenditure related to these applications will reach nearly $12 billion by 2020
[Linask 2018]. Apart from cryptocurrencies, blockchain can also be used in diverse applications,
such as the Internet of Things (IoT).
IoT applications allow direct communications and interactions between devices over the Internet.
Current IoT devices comprise smartphones, smart home appliances, vehicles, and indoor and
outdoor sensors. Before 2020, the number of IoT devices will grow to 20 billion [Jackson 2018].
However, with the involvement of a huge number of devices, traditional IoT applications are facing
challenges in many aspects, including data integrity, security, and robustness.
Blockchain provides a practical solution to address many of the limitations of traditional IoT
applications. Blockchain can ensure IoT data integrity without a third party, while saving bandwidth
and computational power of IoT Devices. Moreover, blockchain can provide a secure and scalable
framework for an IoT network so that sensitive information can be delivered without a centralized
server. Equipped with identification and authentication functions, blockchain can track sensor
data measurements and transfer data among IoT peers without a central server. Furthermore,
blockchain can reduce the operational cost of the IoT. By allowing smart devices to make automated
micro-transactions, two parties can reach an agreement much faster and at a lower cost. A surge in
numbers of applications combining blockchain and the IoT can be seen, with nearly 20% of IoT
deployments adopting blockchain services by 2019 [Bieler 2018]. For example, IOTA [Sikorski et al.
2017], one of the world’s major cryptocurrencies, enables micropayment transactions between IoT
devices.
However, it is still challenging to deploy IoT applications on blockchain systems. First, the
architecture of an IoT-blockchain system needs to support an enormous number of IoT devices.
Second, the consensus, a mechanism to ensure data integrity among peers in the blockchain,
needs to be specifically designed for IoT blockchain due to the limited storage and the computing
capability of IoT devices. Third, traffic modeling of a blockchain network is needed to realize high
system performance in IoT blockchains. A deep understanding of a traffic model can reveal clues
for optimizing communication processes and protocols.
Some previous work has discussed the integration of IoT and blockchain. In [Christidis and
Devetsikiotis 2016], smart contracts of IoT blockchain are investigated and certain unsolved issues
are pointed out. It concludes that the IoT-blockchain combinations can lead the way for new
business models and applications. Marco et al. conduct a systematic literature review of 18 use cases
of the blockchain and investigated system integrity, anonymity, and adaptability [Conoscenti et al.
2016]. Alfonso et al. categorize IoT blockchains into different domains and survey their challenges
[Panarello et al. 2018]. A study of a smart home presents the efficiencies of the integration of
blockchain with home devices [Dorri et al. 2017a].
Different from previous work that provides surveys of IoT networks [Da Xu et al. 2014; Whitmore
et al. 2015] and blockchain systems [Zheng et al. 2018] separately, a survey of IoT applications
using blockchain technology is presented in this paper. This paper also provides a survey of traffic
analysis of IoT blockchain systems, which has not been done before. With a solid survey of many
IoT blockchains, we are in a new position to present several key observations and propose new
architecture and traffic models. Specifically, we delve into three critical aspects of IoT blockchain,
i.e., architecture, consensus and traffic modeling.

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 3

To summarize, this paper makes the following contributions.

• We survey IoT-blockchain systems, and analyze their architectures. We also propose a general
IoT-blockchain architecture.
• We make comparisons between current consensus algorithms and communication protocols,
and show their strengths and shortcomings when applied in IoT blockchains.
• We analyze the current traffic models of P2P and blockchain systems. Moreover, we propose
a traffic model of IoT-blockchain systems.
The rest of this article is structured as follows. Section II gives the background of blockchains
and IoT systems. In Section III, we survey the architectures of IoT blockchains. The consensus
mechanisms are discussed in Section IV. The traffic model analysis is presented in Section V.

2 BACKGROUND
2.1 Blockchain
2.1.1 Overview. As the name suggests, blockchain can be considered as blocks chained together.
Since the advent of a novel type of cryptocurrency (e.g., Bitcoin), it is also referred to in public as
digital ledgers. Blockchain is a comprehensively distributed architecture that emphasizes features
such as data consistency, transparency, user privacy, resistance to retroactive alteration etc. Unlike
other centralized systems, blockchain-based systems use a decentralized network, typically a peer-
to-peer (P2P) network [Schollmeier 2001], to distribute data processing tasks to different nodes. By
using a mechanism called consensus, information stored and data generated in each node can be
synchronized.
In the traditional data sharing system, user information and privacy are often revealed, mishan-
dled, stole, and accessed without authorization [Chen and Zhao 2012; Liu et al. 2018; Sun et al. 2018].
Blockchain technology improves data protection by eliminates the need for the central server. It
uses the peer-to-peer network and consensus protocol to verify user data. No single node can act
as a gatekeeper to control others’ data. Also, blockchain improves data ownership by allowing user
owners to define who can access their files. It uses asymmetric cryptography to ensure the safety
of user data. Moreover, the zero-knowledge proof [Ben-Sasson et al. 2013] is used in blockchain to
provide advanced data privacy of data sharing in blockchain.
Blockchains can be mainly categorized as public and private blockchains [Zheng et al. 2017].
Public blockchains [Nakamoto 2008; Wood 2014] allow anyone to participate in the blockchain
network. The participants help the public blockchain to verify transactions and provide truthful
services. Public blockchains are commonly used for cryptocurrencies. Private blockchains [Brown
et al. 2016; Pongnumkul et al. 2017] only allow authentic nodes to join the blockchain network.
The owner of the private blockchain can modify and delete entities on the blockchain. Private
blockchains have been proposed for various kinds of business applications. Consortium blockchains
and permissioned blockchains are also categorized as private blockchains. In consortium blockchains
[Kang et al. 2017; Li et al. 2017], every participant in the blockchain network is a member of the
same group, company, or industry. Permissioned blockchains like Hyperledger Fabric [Cachin
2016], which give special authority to certain nodes, have been proposed for business use in many
cases. Apart from this, there are Hybrid blockchains [Ateniese et al. 2018] that combine the benefits
of public blockchains and private blockchains. They protect user data by a private blockchain
manager and use the consensus of public nodes to validate transactions. Hybrid blockchains are
commonly used in IoT system and supply chain for data protection and access control.
Apart from the differences in access permissions, the enabling components of blockchains,
such as consensus algorithms and communication protocols, can also be different. For consensus
algorithms, public blockchain system Bitcoin adopts Proof-of-Work (PoW) [Nakamoto 2008], while

, Vol. 1, No. 1, Article . Publication date: November 2019.

4 L.Lao et al.

other public blockchain systems, such as Peercoin and ShadowCash, choose Proof-of-Stake (PoS)
[King and Nadal 2012]. Consensus algorithms, such as Byzantine Fault Tolerant (BFT) [Vukolić
2015], are usually used in permissioned blockchains. Many other algorithms, such as PoA, RBFT in
consensus can also be found in different blockchains.
Gossip and Kademlia are two frequently used communication protocols chosen to support the
message transfer in different nodes. Gossip, which is used in Bitcoin, mimics the spread of epidemic
diseases by transmitting information to the whole network with every node communicating only
with its neighbors. Kademlia, on the other hand, uses a distributed hashtable (DHT) to specify the
network structure and the list of peers every node should communicate with.
2.1.2 Blockchain System Structure. The system structures of Bitcoin and Ethereum are representa-
tive of the public blockchains. The Bitcoin network has a hybrid network architecture containing
supernodes and common nodes. The nodes in the network have four main functions: wallet, mining,
block database, and network routing [Antonopoulos 2014]. Network routing is shared by all nodes.
Other functions are specialized and different from node to node. Some nodes may contain only a
part of these functions. The nodes that contain all functions are called core nodes.
Every node will participate in validating and broadcasting the transaction and block information,
and discovering and maintaining a connection with other nodes. Nodes containing all of the
transaction and block information are called full nodes. Other nodes, containing part of the block
database information, are called light nodes. The wallet node is usually a PC or mobile application
designed to manage users’ accounts and initiate transactions.
2.1.3 Applications. Commercially, blockchain has been proved to be a successful practice, such as
the most famous blockchain system Bitcoin [Nakamoto 2008], which has revolutionized today’s
digital trading system. With the underlying blockchain technology, there are also other applications
in the financial field. For example, Alipay (HK) users can send money to the Philippine wallet Gcash
using the blockchain technique [Chen 2018].
Many famous projects based on blockchains, including Ethereum [Wood 2014], Hyperledger
[Androulaki et al. 2018] and IOTA [IOTA 2019], are under active development. Some advanced
functions, such as smart contract provided by Ethereum [Wood 2014], have also contributed to
accelerate the advent of large decentralized applications [Suryanarayana and Taylor 2004], such
as decentralized medical claim systems [Ekblaw et al. 2016] and decentralized voting systems
[Pilkington 2016]. Apart from its success in commerce, blockchain can also be used in lots of other
fields, such as the Internet of Things (IoT).

2.2 IoT
2.2.1 Overview. With the boom in wireless technologies, Internet of Things applications have
become more pervasive than ever before. The Internet of Things can be characterized as physical
devices with network functions, computers, and items embedded with connectivity functions
[Atzori et al. 2010]. Together with rapid advancements in software and hardware, IoT related
technologies have already become indispensable in today’s society. While many IoT applications
were initially designed for specialized use, most relevant technologies have been migrated and
adapted for public use [Li et al. 2011a].
2.2.2 Applications. IoT applications can be found in many areas for different purposes. For example,
a large number of electronics companies are designing devices for smart homes [Stojkoska and
Trivodaliev 2017]. Xiaomi, a popular electronics company, has already released nearly 100 different
electronic products for smart homes. The IoT is also used in smart cities to provide more conve-
nience to the public [Zanella et al. 2014]. Shared bicycles and shared power banks are deployed in

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 5

many places, e.g., almost all first-tier cities in China. IoT technologies are also applied to public
transportation. Electronic tickets (or E-tickets) can save passengers a great deal of time [Xie and
Shugan 2001]. Mobile payments provide much more convenience, as an effective replacement for
cash.

2.3 IoT Blockchain

2.3.1 Overview. IoT blockchains are blockchain systems that are customized and optimized to
enable IoT applications. IoT applications have been developed and applied in many fields [Miorandi
et al. 2012]. However, most of these applications are prone to problems, such as data leakage and
systematic failure. To mitigate these problematic effects, blockchain can be used to provide higher
security and stability for traditional IoT applications.
Despite the benefit that blockchain brings to traditional IoT applications, there are still many
barriers in its actual implementation. The major problems come from the limitations of IoT devices.
Issues such as task distribution, power consumption, and computational ability need to be taken
into account before blockchain can be effectively applied in most IoT applications. To cope with
these issues, many attempts to adopt blockchain in IoT applications have been seen in recent
years [Bahga and Madisetti 2016; Huh et al. 2017; Sharma et al. 2017]. Although IoT blockchain
applications are in many aspects far from mature, we are optimistic that blockchain can hopefully
confer more capabilities and opportunities in the IoT world with an increase in real-world practices
and studies on IoT blockchain.

2.3.2 Applications. Many IoT applications now adopt blockchain for various purposes [Bahga and
Madisetti 2016; Huh et al. 2017; Sharma et al. 2017]. Generally, there are three major categories,
including digital payment, smart contract service, and data storage.
Digital Payment Digital payment is the first and most widely used field for blockchain [Swan
2015]. While it initially works on a distributed network supported by mostly high performance
machines, special optimization now supported by major blockchains such as Bitcoin and Ethereum
is used for devices with trivial computation power, such as smartphones and pocket PCs. Instead of
being assigned massive computational work, low-end devices usually work as light nodes, which
do not keep the complete chain in their local storage or participate in the most power consuming
processes, like mining. These features have made mobile payment with the blockchain technique
much more accessible than before.
Smart Contract The smart contract serves as a mechanism to make blockchain systems flexible
and scalable to cope with contract-related tasks [Christidis and Devetsikiotis 2016]. In a narrow
sense, smart contracts are considered rules or agreements implemented in a program to automate
transactions. Broadly speaking, smart contracts are viewed as general procedures to build automa-
tion systems and control IoT devices. Currently, many companies provide IoT solutions using a
smart contract [Buterin et al. 2014]. For example, LeewayHertz [LeewayHertz 2019] is a company
that provides solutions for IoT startups and enterprises using Ethereum smart contract. A drug
supply chain network on blockchain is developed by TraceRx [TraceRx 2019]. Similar projects have
been launched in companies like Techracers[Teachracers 2019], Attores[Attores 2019], etc.
Data Storage Blockchain can be utilized as a distributed secure database in data storage appli-
cations. Project DokChain [DokChain 2019] provides a viable solution in healthcare improvement.
It uses permissioned chain (Hyperledger Sawtooth) for protecting the privacy of customer’s data.
Also, it uses blockchain as a distributed database to group financial and clinical data to simplify
the treatment process. Factom [Factom 2019] is a company providing blockchain capabilities for
traditional applications. It provides APIs to integrate data into its blockchain to ensure data integrity.
The blockchain data storage also needs to consider the data authenticity when a user queries a

, Vol. 1, No. 1, Article . Publication date: November 2019.

6 L.Lao et al.

blockchain system. A middleware solution is given in [Peng et al. 2019] to ensure both the query
efficiency and data authenticity.

2.3.3 Challenges. Despite the benefits of IoT blockchain, there are some challenges of adopting
blockchain technology in IoT system [Atlam et al. 2018; Dorri et al. 2016; M.Padma et al. 2019;
Ramachandran and Krishnamachari 2018; Reyna et al. 2018]. They can be summarized into two
categories.
Resource Constraints
General blockchain requires high computational power, high bandwidth, and low delays. Most
blockchain systems use Proof-of-Work (PoW) as their underlying consensus mechanism. However,
the mining process in PoW requires huge computational power. Most IoT devices have simple
hardware specifications and low processing power. It is not capable or time consuming for IoT
devices to perform the mining tasks of blockchain.
Apart from this, blockchain needs to perform data encryption frequently. However, the encryption
speed and time will be different because different IoT devices have different computational power.
Moreover, other processes, consistency algorithms, and routine testing require huge processing
power, which overloads the low power capacity of IoT devices.
Moreover, in blockchain, the transactions and blocks are not stored in a central server. But some
nodes need to keep a copy of the full ledger in their storage. The size of the ledger will increase
over time. However, the majority of IoT devices have low hardware storage capacity. Low power
IoT devices only have 10KB to 100KB memory for storing data and memory [Bormann et al. 2014].
But blockchain requires massive storage for storing the entire chain. For example, Bitcoin needs
over 200 GB of memory, Etherum requires over 1.5 T of memory. It is not capable of storing a copy
of the full blockchain for IoT components.
Additionally, the consensus process in blockchain requires the exchange of information between
nodes frequently to reach an agreement to maintain the correctness of blockchain and generate
new blocks. This process requires high bandwidth and low network latency. However, IoT devices
are always strict in a limited bandwidth.
Scalability
According to the blockchain trilemma by Vitalik Buterin, it is hard to achieve high scalability in a
secure and fully decentralized blockchain network. However, the IoT system is expected to involve
numerous IoT devices. Also, IoT devices always require to join and leave the network frequently in
many application scenarios. For example, PBFT consensus mechanism is considered as a suitable
protocol for IoT system. However, the PBFT algorithm can only work well in a fixed-size network,
in which the network members cannot easily change. That is not scalable for the numerous IoT
devices management.

2.4 Current Trend in IoT Blockchain Development

With the rapid growth of IoT and blockchain technology, the development trend of IoT blockchains
also continues to change. The current trend is mainly reflected in four aspects: popularity, range of
applications, development of underlying technology, and business models.

2.4.1 Popularity. In the past 10 years, the IoT and blockchain industries both grew rapidly [Whit-
more et al. 2015; Zheng et al. 2017]. In the beginning, IoT was only applied in certain fields, such as
manufacturing, logistics, and transportation. Then, more and more emerging industries became
involved in the wave of IoT, such as smart homes, smart cities, and asset digitalization. A growing
trend of research in related fields also indicates a growing interest in academia [Khan and Salah
2018; Kshetri 2017; Zheng et al. 2018]. Different kinds of consensus algorithms have been presented

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 7

and structural improvements have been proposed for IoT blockchains [Samaniego and Deters 2016;
Zhang and Wen 2017].
2.4.2 Range of Applications. IoT blockchain has also become more widely applied than ever before.
The initial uses of blockchain have focused on decentralized currency systems [Gervais et al. 2014].
One of the most famous blockchain systems, Bitcoin, was first designed to create a decentralized
currency system without any supervision. With the development of blockchain technologies,
blockchain systems with more advanced features, such as smart contract in Ethereum, have enabled
a wider variety of applications and services other than financial use. The integration of IoT into
blockchain systems provides even more possibilities for applications. Logistics companies such as
SF Express focus on using blockchain to implement asset tracking. Hardware and electric product
companies, e.g. MI, will use blockchain to enhance the interaction between humans and IoT devices.
Energy companies use blockchain to implement energy sharing and energy transaction business.
IoT blockchains can also be deployed in the edge computing environment where IoT devices contain
limited resources and a resource allocation strategy is proposed in [Huang et al. 2019].
2.4.3 Development of Underlying Technology. The rapid development of many underlying tech-
nologies also contributes to the rapid growth of IoT and blockchain. In view of the scale of IoT
connectivity, with the development of communication technologies such as LoRa, NB-IoT [Sinha
et al. 2017] and 5G [Andrews et al. 2014], communications using IoT devices in China has increased
rapidly. By the end of June 2018, the number of Chinese Internet accesses reached 465 million,
twice the amount of the previous year. To meet different blockchain application requirements,
newly-proposed structures are equipped with optimization functions. For example, many new
blockchain systems adopt fast consensus algorithms for quick transaction confirmation. Public
chains like IOTA [Sikorski et al. 2017] and EOS [Kakushadze and Russo Jr 2018] solve the low
transaction rate problem of traditional implementations. Private chains like Hyperledger [Cachin
2016] are more suitable for consortium business.
2.4.4 Business Models. Many companies are seeking ways to integrate blockchain techniques with
their business models to enhance profits. For example, the E-Business company, JD [JDChain 2019],
deploys a blockchain system to solve problems occurring in logistics and electronic invoices. There
are also several start-up companies, such as Slock.it [sloct.it 2018], which use existing blockchain
systems such as Ethereum to implement P2P business. In this way, IoT devices will connect to
the current blockchain system and users can use smart contract to interact with these devices.
However, the blockchain system performance may severely affect the supported business, e.g., how
fast the consensus algorithm validates transactions, which deserves more attention when designing
industrial applications.

3 ARCHITECTURE
This section presents the architecture of IoT, blockchain, and blockchain-based IoT. In addition, we
investigate the current IoT blockchain applications, give a general architecture of IoT blockchain,
and conduct several case studies.

3.1 IoT Network Architecture

The traditional IoT architectures are well discussed in previous work [Al-Fuqaha et al. 2015; Da Xu
et al. 2014; Said and Masud 2013; Whitmore et al. 2015]. From the pool of proposed architectures,
there are two types of IoT architecture: three-layer architecture [Al-Fuqaha et al. 2015; Khan et al.
2012; Said and Masud 2013; Wu et al. 2010; Yang et al. 2011] and five-layer architecture [Al-Fuqaha
et al. 2015; Atzori et al. 2010; Chaqfeh and Mohamed 2012; Said and Masud 2013; Tan and Wang

, Vol. 1, No. 1, Article . Publication date: November 2019.

8 L.Lao et al.

2010; Wu et al. 2010]. The three-layer architecture, which is illustrated in Fig. 1 (a), comprises three
basic layer.
Physical layer: The physical layer consists of a variety of IoT devices including sensors, RFID
tags, NFC devices, and mobile phones. This layer is responsible for connecting different devices,
exchanging messages, and gathering information to the upper layer.
The mobile phone is an integrated IoT device which includes many IoT elements, including
camera, microphone, proximity sensor, ambient light sensor, accelerometer, gyroscope, magnetic
field sensor, temperature sensor, RFID, and NFC. There are billions of mobile phone users gener-
ate massive data at every single moment around the world. Social network services and social
applications rely on mobile phones. User information like messages, voices, photos, and location
information is collected by mobile phones and uploaded onto social platforms.
Radio-Frequency Identification (RFID) is a wireless radio communication technology. RFID
systems exchange data between RFID tags and RFID readers by electromagnetic fields. Each RFID
tag contains a unique identifier, which allows RFID readers to distinguish it from others. The RFID
technology is widely used in the indoor pedestrian localization, indoor object positioning, and
commodity tracking in logistics systems. The IoT system employs the RFID technology for physical
security, device management, and monitoring service.
Near Field Communication (NFC) is a communication protocol which allows devices to com-
municate within short ranges. NFC provides an easy communication channel and data sharing
approach without the need to set up a network. Also, NFC technology can be utilized to connect
unpowered objects, which provides a more flexible approach to accessing IoT devices around us.
As a large amount of NFC-enabled devices are deployed, NFC technology is playing an essential
role in the IoT world.
Other IoT hardware devices include various kinds of wireless sensors for environmental monitor-
ing. A large wireless sensor network can consist of numerous wireless sensors. They work together
to perform tasks like data collection, data monitoring, and data analysis in wide ranges. Common
wireless sensors include infrared sensors, temperature sensors, ultrasonic sensors, pressure sen-
sors, chemical sensors, and smoke sensors. They perform the environmental monitoring of light,
temperature, sound pressure, food quality, and air quality.
Network layer: The network layer transfers information which collected or processed by IoT
devices. Data is conveyed through diverse technologies such as 4G, 5G, RFID, WiFi, and Bluetooth.
Specifically, the network layer connects edge nodes, hypervisors, and user applications.
Application layer: The application layer can be recognized as an abstraction of IoT services
provided to users. A user can request different services through APIs. After receiving user requests,
the application layer either processes the information locally or calls the underlying APIs to handle
the requests.
With the rapid development of the IoT, many auxiliaries and business models have emerged.
The rudimentary three-layer architecture model cannot provide a sufficiently accurate abstraction.
Accordingly, a more concrete five-layer architecture is proposed to fulfill the description of IoT,
which is exhibited in Fig. 1 (b).
In the proposed five-layer architecture, the physical layer is renamed the sensing layer, because
of the widely-deployed sensors in IoT systems. In addition, the network layer was expanded to the
network layer together with the middleware layer. The new network layer still undertakes the task
of data transmission while the middleware layer is responsible for data processing, data storage,
and service management.
Middleware layer: Since there are numerous heterogeneous devices in the lower layer of the
IoT system. They may have different operating systems, query formats, and data formats. It is
difficult for software developers to develop an integrated IoT application which is compatible

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 9

with all IoT devices. Middleware layer sits between IoT devices and IoT applications to handle the
compatibility problem. It handles users’ requests and gives efficient responses. The functionalities
provided by the middleware layer consist of primary message delivery, API management, data
and service integration, and data formatting. Network companies like Cisco [Cisco IoT 2019],
Huawei [HUAWEI IOT 2019], Oracle [Oracle IoT 2019], and Mulesoft [MuleSoft IoT 2019] offer IoT
middlewares.
Moreover, the business layer was added to the top of application layer.
Business layer: As the IoT develops, the applications supported by an IoT system have increased
dramatically, forming a complicated ecosystem. A single application layer cannot accurately describe
it. Therefore, the business layer was proposed to represent a higher-level abstraction of the IoT
ecosystem. It is responsible for the management of the entire IoT system. Also, the business layer
builds business models, graphs, and performs data analysis.
Depended on the analysis results, the information provided in this layer offers guidance and
recommendations to help system managers in decision-making and future development.

Business Layer
Application Application
Layer Layer
Application
Layer
Propagation
Layer
Middleware
Layer
Consensus
Network Layer
Layer
Network Layer
Network Layer
Physical Layer Sensing Layer
Physical Layer
(a) (b)

Fig. 1. IoT general architecture: (a) three-layer archi- Fig. 2. Five-layer architecture of blockchain.
tecture; (b) five-layer architecture.

3.2 Blockchain Architecture

The architectures of blockchain have been broadly discussed in a previous work [Lin and Liao
2017; Swan 2015; Zheng et al. 2018]. We propose a five-layer architecture model, which summarizes
previous discussions. This model is illustrated in Fig. 2.
The physical layer consists of various kinds of blockchain nodes. They interconnect with each
other to form the blockchain network. In general, there are two types of nodes in blockchain. Full
nodes and light nodes.
Full nodes: Full nodes keep a copy of the full blockchain in their storage. They participate
in creating, verifying, and adding new blocks to blockchain. Accordingly, full nodes maintain
the security, consensus, and data correctness on blockchain. Network participants require high
hardware storage to become a full node. Some full nodes can be miners that contribute their
computation power and resources to get a relative amount of reward. Miner completes a hash
calculation and broadcasts his result to the blockchain network. Then, the result will be verified by
others under a consensus protocol. Once the result is confirmed, the miner gets his reward.

, Vol. 1, No. 1, Article . Publication date: November 2019.

10 L.Lao et al.

Light nodes: Light nodes don’t require high hardware specifications. They don’t keep a copy of
the full blockchain but just know the status of the last block. Light nodes obtain the blockchain
information from full nodes.
The network layer provides reliable data delivery service. It allows the blockchain nodes to
communicate with each other. In general, blockchain network is based on Peer to Peer (P2P)
network. In the network, there are no central servers and central authority. Network participants
reach an agreement to ensure the integrity and security of the network under a consensus protocol.
Accordingly, blockchain achieves decentralization and avoid centralized vulnerabilities.
The consensus layer is responsible for ensuring the consistency of the stored data and incentivizes
participants to find a new block. The consensus mechanism is the important part in blockchain to
ensure data correctness and integrity. We discussed the consensus mechanisms in depth in Section
IV.
The propagation layer consists of communication protocols that define the rules of how messages
[Wen et al. 2014] and blocks [Decker and Wattenhofer 2013] are propagated in the network. Bitcoin
and Hyperledger fabric adopt the Gossip as their communication protocol while Ethereum employs
the Kademlia algorithm as the basis of its communication protocol.
Gossip protocol: Gossip protocol represents a kind of epidemic-like peer-to-peer communica-
tion method. In a gossip-based system, updates spread like an infectious disease and eventually
infect nearly all nodes. Therefore, gossip protocol utilizes the terminology of epidemiology to
denote different nodes types. A node which owns an update and is about to spread to others is called
the "infective" node. In contrast, a node which waits to accept an update from others is denoted
as the "susceptible" node. Specifically, a node that owns an update but is not going to spread it is
called the "removed" node.
Gossip protocol has two different modes of propagation: anti-entropy and rumor-mongering.
The anti-entropy method reduces the entropy of the system by eliminating the differences between
distinct nodes and making the whole system orderly. Specifically, a node with an update periodically
chooses one or several nodes at random and compares contents. If discrepancy of contents is found
among the nodes, they will resolve the differences through data-pulling or data-pushing to get
each one updated. Rumor-mongering adopts a different method. It merely propagates the updates
to a random node on a regular basis. The whole system can achieve eventual consistency as time
goes by.
The two modes of propagation require different types of nodes. In anti-entropy, only susceptible
and infective nodes are needed. Therefore, anti-entropy is also recognized as the SI model or simple
epidemics model. In contrast, rumor-mongering needs nodes of all the three types and is recognized
as the SIR model or complex epidemics model.
Kademlia: Kademlia is an algorithm designed for decentralized peer-to-peer networks. It is
used in Ethereum to optimize network routing and to help locate target nodes. In Kademlia, each
node has its unique node ID. The Kademlia algorithm uses node ID to perform fast peer lookup
and locate files or resources.
In the Kademlia network, each node is recognized as a leaf on a binary tree. The logical distance
between two node IDs can be calculated by applying the XOR operation. The smaller result
represents the closer logical distance between two nodes. The nodes with the longest common
prefix (LCP) will share a subtree. K-buckets can be expressed as subtrees. The K in K-bucket refers to
the maximum number of leaves in the subtree. The routing table in Kademlia stores information in
a K-bucket list. Fig. 3 gives an example of the binary tree, routing table, and K-bucket in Kademlia.
The process of peer lookup in Kademlia is illustrated in Fig. 3. Assume the node 001 hopes to
locate a target node 101 (find the IP address of node 101). The Kademlia algorithm will first calculate
the logical distance of 001 and 101 by XOR operation and get operation result 100, which means

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 11

bucket0 000 LCP 00

1 0 Peer
bucket1 010 011 LCP 0
001
bucket2 110 100 111 No LCP

1 0 1 0 bucket0 111 LCP 11

Peer
bucket1 100 LCP 1
110
bucket2 011 010 001 000 No LCP
1 0 1 0 1 0 1 0
101 bucket0 101 LCP 10
Peer
bucket1 111 110 LCP 1
100
110 100 001 bucket2 001 100 010 011 No LCP

Fig. 3. Illustration of peer lookup in Kademlia.

Table 1. Summary of Blockchain Applications

Blockchain Application Consensus Mechanism TPS (tx/sec) Releases date

Bitcoin PoW 7 Jan 2009
Litecoin PoW 56 Oct 2011
Bitshares DPoS 17 Jul 2014
NEO DBFT 1000 Feb 2014
Ethereum Pow/PoS 15 Jul 2015
Hashgraph Hashgraph 10000 Jul 2017
Tangle DAG 800 Apr 2018
Ripple Ripple 1700 May 2018
EOS DPoS 3000 Jun 2018
QTUM PoS 70 Jul 2018
Futurepia DDPoS 300000 Sep 2018
Casper PoS 10000 Jan 2019
Monoxide PoW with Chu-ko-nu Mining 15.6 Feb 2019

the largest common prefix of the IDs is 1. The algorithm will then try to find the target node from
2-bucket(K = 3 − 1 = 2) in the routing table of peer 001. If the target node is found in 2-bucket,
then this process finishes.If not, the algorithm will send a lookup request to 110 (The first record in
2-bucket). Then the lookup operation will be performed recursively until the target node 101 is
found.
The application layer is responsible for implementing functions according to the application
requirement. Blockchain applications have been widely used as cryptocurrency in financial markets.
Also, there is a clear trend that blockchain applications will internationally adopt in health care,
IoT, and supply chain shortly. We summarize and compare the consensus mechanism, performance
in transaction per second, and releases date of representative blockchain applications in Table 1.

, Vol. 1, No. 1, Article . Publication date: November 2019.

12 L.Lao et al.

3.3 General IoT-Blockchain Architecture

After illustrating the typical architectures of IoT and blockchain, we conclude a five-layer general
architecture of IoT-blockchain from the existing IoT blockchains. It is given in Fig. 4. The architecture
of existing IoT blockchains are summarized in Table 3. This architecture combines the features
of traditional IoT systems and blockchain systems. The physical layer of IoT blockchain is the
same as the physical layer of IoT [Gubbi et al. 2013; Lee and Lee 2015]. It includes the sensors,
smart home devices, RFID tags, mobile phones, and other IoT devices that related to the IoT-
blockchain application. The network layer of IoT blockchain is similar to the network layer of
the traditional blockchain layer. It is responsible for internetworking, routing, and multicasting
between IoT devices [Jiang et al. 2016; Maymounkov and Mazieres 2002; Milojicic et al. 2002]. In
our survey, the majority of IoT-blockchain applications are based on the P2P network. Few of
them uses the commercial network. The blockchain layer represents the blockchain functions in
IoT blockchain, such as consensus, data storage, and data sharing. It can be a specific blockchain
platform, commercial blockchain, or public blockchain solution [Crosby et al. 2016; Nakamoto
2008; Wood 2014]. The middleware layer of IoT blockchain is different from the middleware layer
of IoT. The main purpose of this layer is not data formatting or integration for heterogeneous
IoT devices. But it responsible for blockchain management, blockchain service integration, and
providing additional security service [Alphand et al. 2018; Bahga and Madisetti 2016; Huh et al.
2017]. The application layer of IoT-blockchain architecture is similar to IoT system and traditional
blockchain architectures. It provides interactions, abstraction services, and API functions to users.
In general, IoT-blockchain application is included as one of the blockchain application. Also, it can
be known as an IoT application that implements blockchain technology [Dorri et al. 2017a; Huckle
et al. 2016].
In this architecture, IoT devices are responsible for data generation. Blockchain can serve as a
secure distributed database that safely keeps records and avoids malicious modification. Once a
block is confirmed and added as a part of the blockchain, the transactions contained in the block as
well as transactions in all previous blocks get tamper-proof. Additionally, all historical records can
be easily retrieved without loss of information. The authorized user can browse all transactions
and verify the authenticity of each transaction. For most applications, blockchain is used as an
add-on storage technique. In addition, blockchain can support cryptocurrency and optimize digital
trading in IoT systems.

3.4 Current IoT-Blockchain Applications and Their Architectures

IoT-blockchain applications are still at an early stage. However, the integration of IoT and blockchain
is rapidly evolving and growing. We summarize and compare the service, blockchain, consensus,
IoT devices, and company size of some typical IoT-blockchain application in Table 2. We also
investigate and present the characteristics and architectures of the representative IoT-blockchain
applications below. The architectures of these applications are similar to the general IoT-blockchain
architecture that we proposed in Fig. 4. They are categorized into two types:
3.4.1 Specific Application. Specific application refers to the standalone software or system that
used IoT and blockchain as an essential component in their commercial operations. The core layers
of this category are the application layer and the physical layer. The IoT devices and the functions
of the application define the nature of the IoT-blockchain application. The blockchain layer and the
network layer are responsible for storage and communication. Some IoT-blockchain applications
with a simple architecture do not include the middleware layer.
Dorri et al. introduce LSB blockchain structure that emphasizes the security and privacy in a
smart home [Dorri et al. 2017b]. The architecture of LSB is shown in Table 3. Physical layer includes

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 13

Application Application interface

Layer IoT APIs

Middleware Service Intergration

Layer

Blockchain
Layer
Decentralized
Blockchain data storage

Network Routing
Internetworking Routing Multicasting
Layer protocols

Data collection protocols

Physical
Layer Mobile
RFID tags NFCs Sensors
Phones

Fig. 4. 5-layer architecture for an IoT blockchain.

various types of Smart Devices. High resource smart devices are responsible for managing a public
blockchain to ensure user privacy and security. Low smart resource IoT devices are responsible for
end-to-end communication and processing entering and outgoing requests. The smart devices are
registered on the blockchain network. Middleware layer provides overlaid blockchain management
that can reduce the management overhead of the blockchain. Smart home applications achieve
home automation.
LO3 Energy [LO3 2018] introduces a P2P marketplace for solar energy. The architecture of LO3 is
shown in Table 3. The physical layer of LO3 includes electrical grid and solar panels. They itemize
their extra energy yield and upload it to blockchain through a low latency network. Exergy token
system is used as a middleware to enhance the system performance and the efficiency of network
participants’ authorization. The customer can purchase energy by commercial energy shopping
applications.
"Slock.it" [Prisco 2016] is an existing IoT-blockchain application in the market. It operates with
electronic locks that can unlock by an appropriate token. The client who wishes to sell their
properties can set a price on an electronic lock. The customer can browse the goods and pay the
requested amount in cryptocurrencies to unlock the lock. The architecture of "Slock.it" is shown in
Table 3. It has a simple architecture that consists of distributed applications, Ethereum blockchain,
commercial network, and electronic locks.
3.4.2 Application Platform as a Service. Application platform as a service refers to the support
software that connects everything in an IoT-blockchain system. It integrates IoT devices and
blockchain technique and provides a straightforward management platform for developers. It

, Vol. 1, No. 1, Article . Publication date: November 2019.

14 L.Lao et al.

facilitates communication, data flow, and device management. The core layer of this category is
the middleware layer. The nature of middleware defines the functions of the platform service.
In this category, the elements of the application layer and the physical layer are not specified.
The blockchain layer and the network layer need to support the functions of the middleware. For
example, allow data integrity check in the smart contract.
Sagirlar et al. propose a new IoT-blockchain platform "Hybrid-IoT" [Sagirlar et al. 2018]. The plat-
form implements consensus based on PoW and BFT algorithms. The authors define sub-blockchains
and inter-blockchains as the area structures of IoT-blockchain. Fig. 5 shows the architecture of
hybrid-IoT. In the architecture, two PoW sub-blockchains are connected by a BFT inter-connector
framework.

PoW sub-blockchain 1 PoW sub-blockchain 2

Transaction Pool 2
Transaction Pool 1

Cloud Service

Outsider Device Full Peer Device Light Peer Device Data Transaction Inter-Chain Transaction

Fig. 5. Architecture of Hybrid-IoT.

Bahga et al. introduce a platform for the industrial IoT (BPIIoT) [Bahga and Madisetti 2016]. This
platform allows users develop applications (DApps) with the features of blockchain. Its architecture
is shown in Table 3. In the platform, IoT devices need to register on blockchain network. Users
can develop applications on single-board computers (SBC) to control and manage the blockchain
network and IoT devices.
"JD.COM" e-business company publishes a blockchain platform "JD Blockchain Open Platform"
[JD 2018] that focuses on providing IoT solutions with the blockchain technology. This blockchain
platform provides blockchain gateway service, blockchain node service, and blockchain consensus
network service. The architecture of JD Blockchain platform is shown in Table 3. The platform uses
BFT-like consensus algorithm. Also, an authentication protocol is used to control the number of
accesses to the blockchain network. The system has three kinds of peers: consensus peers, gateway
peers, and IoT devices. Gateway peers function in the middleware layer to integrate inputs and
protocols from lower layers.
Liu et al. [Liu et al. 2017] propose a framework to implement IoT data integrity and security based
on a blockchain system. The authors describe critical components of the proposed decentralized
data integrity service framework. The framework achieves full decentralization with a data integrity
verification protocol. It provides reliable data integrity verification to the users in an IoT system
without requiring a third-party auditor. The architecture of this framework is shown in Table 3. In
the framework, the IoT device is responsible for generating data and writing data to Ethereum. The
user can check the data integrity provided by data integrity service through a data user application.
Alphand et al. [Alphand et al. 2018] propose an architecture "IoTChain" which combines OSCAR
architecture [Vučinić et al. 2015] and ACE authorization framework [Seitz et al. 2017]. In IoTChain,
each registered user has an authorized token. It identifies the particular privilege of a set of resources.
When a user wishes to access an object, the user needs to send a transaction with the required
data to the smart contract address. Then the smart contract will generate an authorized token
for the data user. This architecture uses blockchain instead of the centralized ACE authorization

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 15

Table 2. Comparison between IoT-Blockchain Applications

IoT-Blockchain Service Blockchain Consensus IoT devices Company size

Transaction Hardware-based Blocklet USB Enclave,
Filament [Filament 2018] PoW 40 milions market cap
service to embedded IoT Consortium Blockchain Blocklet Chip
Fabric Broker,
Xage [Xage 2018] Security service Fabric 300 milions market cap
consensus Enforcement Point
Integrated Sensors,
UniquID [UniquID 2018] Litecoin PoW Open source project
service to IoT and blockchain Actuators, Appliances
Public Robots, More
LeewayHertz [LeewayHertz 2019] IoT-blockchain solutions PoW
blockchain Audio devices than 10 years in operations
ElectriCChain [ElectriCChain 2018] Process data of solar panel SolarCoin PoS Solar panel Open source project
Atonomi Smart devices, Leading provider
Atonomi [Atonomi 2019] IoT-blockchain solutions Atonomi
consensus Smart home of IoT data security
Public Grid Edge,
LO3 Energy [LO3 2018] Solar energy marketplace PoW 1 million in revenue annually
blockchain solution Solar plane
Slock.it [sloct.it 2018] Commission shop Ethereum PoW Electronic lock 1.5 millions in revenue annually
JD.com [JDChain 2019] Blockchain platform BFT blockchain BFT IoT devices 1.7 trillions market cap

Table 3. Summary of IoT-Blockchain Architectures

IoT Blockchain Application Layer Middleware Layer Blockchain Layer Network Layer Physical Layer
Overlay blockchain
Smart home Smart home application Commercial blockchain P2P network Smart device
management
Energy shopping Public blockchain
LO3 Energy Exergy token system low latency network Grid Edge, Solar plane
application solution
Slock.it DApp None Ethereum Commercial network Electronic locks
PoW blockchain, Full peer, Light peer,
Hybrid-IoT IoT application Hybrid-IoT platform P2P network
BFT blockchain Sensor
Blockchain network
BPIIoT Manufacturing DApps Single-board computers P2P network Industrial IoT device
bridge
Blockchain gateway
JD.COM JD.com BFT blockchain P2P network IoT devices
service
IoT data Service Data integrity
Data user application Ethereum P2P network IoT devices
Framework service framework
IoTChain Authorized access OSCAR, ACE framework Ethereum Commercial network IoT devices

server. The architecture of IoTChain is shown in Table 3. In this architecture, the IoT device is
responsible for data generation. The data owner is responsible for uploading the data to blockchain.
The OSCAR architecture and ACE authorization framework are responsible for ensuring the safety
of user data.

4 CONSENSUS
The consensus mechanism helps multiple participants in a network to reach a necessary agreement.
Consensus mechanisms should be fault-tolerant in providing reliable services. In the decentralized
environment of blockchain, a consensus mechanism is a set of rules that hinges on the contribu-
tions of multiple participants to collectively maintain a synchronized state. A superior consensus
mechanism should achieve validity and consistency to function appropriately, while efficiency and
cost-effectiveness need to be maintained to ensure high performance.

4.1 Common Consensus Mechanisms

In the blockchain, a consensus mechanism is necessary to provide truthful services among partici-
pants without mutual trust. To keep locally stored ledgers consistent with others, various consensus
mechanisms are implemented for different scenarios. We divide them into four categories based on
their implementations.
4.1.1 Byzantine Fault Tolerance Series. Byzantine Fault Tolerance (BFT) [Driscoll et al. 2003] is a
solution to solve the problem of reaching agreement in a system where processors can fail in an
arbitrary manner, named Byzantine General Problem [Lamport et al. 1982]. BFT is a replica-based

, Vol. 1, No. 1, Article . Publication date: November 2019.

16 L.Lao et al.

approach that exploits communications between replicas to achieve an agreement on request

without restrictions or presumptions.
The Practical Byzantine Fault Tolerance (PBFT) [Castro et al. 1999] is based on the BFT mechanism.
It improves system performance by reducing the complexity of Classic Byzantine General Problem to
polynomial level and enabling it to function correctly in asynchronous systems [Castro et al. 1999].
PBFT network should consist of at minimum 3f + 1 nodes in order to tolerate f defective nodes. In
a blockchain system, it requires 2f + 1 nodes to gain the consensus of the block of transactions.
HQ replication [Cowling et al. 2006] is a protocol that optimized the PBFT. It reduces the amount
of information disseminated, making PBFT more efficient. HQ replication does not fundamentally
change the structure of PBFT, mainly in the case where the client does not become involved in a
competition when sending a request. It classifies the operations as read and writes. Requests from
clients are normally write operations. In process of write, it is divided into two phases: Phase 1,
the client first sends write request to a node, collecting the status information of the server at the
same time. If 2f + 1 nodes are in the same state and agree to execute the client’s request, the client
initiates the second phase, phase 2, causing the node to execute the client request; otherwise, it
means the request encounters the competition. In this case, nodes will execute PBFT to reach a
consensus, then the leader node will finally determine the execution order of the request.
Specular Byzantine Fault Tolerance protocol (SBFT) is also a vital part of BFT protocols. A typical
SBFT protocol is introduced in Zyzzyva [Kotla et al. 2007], which significantly improves PBFT
efficiency. The main idea is that nodes in a Byzantine system are in a normal state most of the
time. Thus, clients need not wait for each request to be executed after consistency is reached. After
receiving client’s request, a node will not perform the costly pairwise interaction as in the PBFT,
but instead directly executes the client’s request. Nodes are only supposed to gain consistency
when an error occurs. In this case, the response time to a request will be greatly accelerated.
Robust Byzantine Fault Tolerance protocol (RBFT) [Corp 2019] is actually a series of protocols that
strengthen the robustness of traditional PBFT. In other BFT protocols mentioned above, researchers
try to enhance the efficiency of the BFT system under normal circumstances. Therefore, when
a Byzantine fault is encountered, the performance of such protocols decreases dramatically. It is
even difficult to ensure the liveness of the system as a whole. Through improvement of the PBFT
algorithm, RBFT ensures that the system performance roughly the same in the best or worst case
conditions, thus providing a truly practical BFT.
However, the BFT series consensus protocol still have some problems. For example, most BFT-
based blockchain systems are permissioned blockchain so that people cannot dynamically join
or exit a blockchain system. Accordingly, it is difficult to deploy a permissionless BFT blockchain
system.

4.1.2 Proof-of-Somethings (PoX) Series. We use Proof-of-Somethings to represent a series of con-

sensus protocols that achieve consensus by demanding the devotion of computational resources
from every node in the entire network. Compared with the BFT protocol and its variations, PoX
algorithms adopt specific probabilistic mechanisms that do not require full information about the
node operations in the system. It is based on the assumption that there are always more benign
nodes than malicious nodes in the network. This means that compared with malicious nodes, benign
nodes always have more resources. Because PoX algorithms need not require all the information
from participant nodes, systems using PoX set no limits for nodes to join or exit.
Proof-of-work (PoW) was proposed by Nakamoto et al. in 2008 [Nakamoto 2008]. It was initially
designed to reduce jam emails and then employed by the Bitcoin. In Bitcoin, the PoW consensus
protocol requires nodes to solve a computationally complicated math puzzle. The mathematical
problem serves as a means to verify the validity of blocks on the network before accepting it.

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 17

Solving this puzzle requires a large amount of computational power for every node. It acts as a
mandatory regulation for every node to obey, thus building trust among all nodes. The blocks
chained using PoW also become extremely resistant to tampering, because of the extremely high
computational cost it requires. However, a blockchain system running PoW may also be attacked in
the mining process where an attakcer can get more profit by launching mining attacks and bribery
attacks [Gao et al. 2019]. A refined version of PoW is introduced in [Dwork and Naor 1992], which
proposes a probabilistic solution to the Byzantine General Problem. It also ensures the stability of
blockchain systems in environments with a large number of untrusted nodes.
Proof-of-Stake (PoS) [Kiayias et al. 2017] relies on a hypothesis that users holding more currency
are more incentived to ensure the reliability of the system, and is less likely to play as malicious
nodes. For pure PoS [Azima 2018], users holding more coins over a long period have a higher
possibility of being selected by the system to generate the next block. In DPoS [Larimer 2014],
people will elect several super nodes according to the amount of currency of each node. News
blocks will be generated by the selected super nodes. This approach decrease the decentralization
of the blockchain system so that it requires new nodes trust these super nodes.
Proof-of-Capacity (PoC) [Dziembowski et al. 2015], also called Proof-of-Space (PoSpace), exploits
the idle space in local computers’ disk for mining. PoC is similar to the PoW consensus protocol,
except that instead of devoting computation resource, space in disks is utilized. In PoC, all the
nodes store the solutions for a series of computation puzzels, which are hard to find but easy to
verify. If a node has a calculation method on its disk, which is the fastest solution corresponding to
the puzzle in the latest block, this node will have the right to issue a new block and get the mining
rewards. One well-known application of the PoC consensus protocol is InterPlanetary File System
(IPFS) [IPFS 2019], a peer-to-peer file system based on the distributed hash table. IPFS provides an
incentive system with the built-in cryptocurrency FileCoin, which can be mined by PoC consensus
protocol. In the IPFS, all the files are given a unique fingerprint, which is a cryptographic hash.
Each network node only stores the content that it is interested in. A user, who wants to find a file,
needs to find the nodes that are storing the required files.
The Proof-of-Authority (PoA) [Naumoff 2016] consensus protocol was proposed to relieve the
requirement of computational power or amount of coins. Instead, POA introduces a new concept
called authority. Different from PoW and PoS, only nodes with authority are allowed to generate new
blocks in systems using PoA and thus ensuring the security of the whole chain. It is more flexible
with less computational overhead compared with other PoX when applied to private systems.
Proof-of-Importance (PoI) [NEM 2018] is a consensus mechanism developed to detemine which
network participants are eligible to add a block to the blockchain. Each account will be evaluated
and given a importance mark. The importance depends on the account’s transaction amount,
transaction parterner number and vested stake. According to the importance score, you can not
only get the corresponding reward, but also do not need to use any computing resources of your
computer, even when you are offline.
Proof-of-Burn (PoB) [Frankenfield 2018] is a consensus protocol that miners should show proof
that they have burned some coins, which is sent these coins to some unspendable addresses. The
amount of these destroyed coins determines the probability of a miners to issue a new block. This
consensus protocol does not need massive useless computation so that it is environmental.

4.1.3 DAG Series. Directed Acyclic Graph (DAG) [Benčić and Žarko 2018] was proposed [Lerner
2015] to improve the parallelism of traditional single-chain blockchain system. DAG employs the
directed graph data structure to connect blocks, and its related consensus mechanism is distinct
from others. Each transaction will be linked with the previous two transaction records. In this case,

, Vol. 1, No. 1, Article . Publication date: November 2019.

18 L.Lao et al.

the legality of the current transaction can be proven by referencing previous transactions. In this
way, the legality of all transactions can be confirmed.
Compared with other PoX consensus protocols, DAG only cares about countable linked transac-
tions. In addition, it is more than simple PoX consensus protocol. Considering the mining process,
DAG employs the PoW to mine a new block. However, DAG blocks are not chained in a linked
list but a tree. Each node only checks the legitimacy of transactions it is associated with. This
mechanism enhances the efficiency and availability with a sacrifice in partition tolerance. Byteball
[Churyumov 2016] introduces a decentralized system that adopts DAG as a consensus mechanism.
In this system, everyone is allowed to add new data into the database by paying a fee proportional
to the data size. Next nodes will confirm the previous data by including their hash into current
data. Moreover, the previous node will receive confirmations from the next nodes.

4.1.4 Ripple Series. There are two kinds of nodes in the Ripple network [Pilkington 2016]: server
node and client node. A Ripple server maintains a Unique Node List (UNL) for the consensus
process, then votes on the authenticity of transactions. If the consensus agreements are more than
80%, then all transactions that recognized legal are included into the blockchain. Then the ledger is
closed as a new last-closed ledger. In contrast to those PBFS nodes that need to query all nodes in
the network, a server in Ripple only queries a minimum percentage of 80% of its UNL.

4.2 Comparison between Consensus Mechanisms

We summarize the characteristics of some widely-deployed consensus mechanisms [Baliga 2017;
Castro et al. 1999; Chalaemwongwan and Kurutach 2018; Cowling et al. 2006; Dziembowski et al.
2015; Frankenfield 2018; King and Nadal 2012; Kozak 2018; Lerner 2015; NEM 2018; Salimitari and
Chatterjee 2018; Zheng et al. 2017] in Table 4, where we compare the blockchain type, throughput,
system scalability, transaction finality, tolerated power of adversary, advantage, disadvantage,
vulnerability, and example of use. From the table, we can see that the consensus mechanism of PoS,
DAG and Ripple have high transaction speed and system scalability. PoS has the best tolerated
power of adversary. However, the tolerated power of DAG will be higher in practice, because there
are coordinator nodes in the system.
In a small scale network, BFT consensus mechanism can provide high throughput and security.
The performance of BFT drops as the number of nodes grows, because of the additional communi-
cation cost between the new and existing nodes. Additionally, BFT has advantages of extensibility
that it can combine with various kinds of improvement algorithms to meet specific needs.
PoX mechanism is suitable for the public chain because the requirement of computational power
or currency is useful in preventing denial of service attacks, service abuse and making the chain
more secure and reliable. However, PoX also suffers from inefficiency and high computational
overhead.
DAG is a new trend in blockchain systems, because it can achieve a high throughput of BFT-like
consensus under the premise of obtaining the flexibility of the PoX series algorithm. It combines the
advantages of BFT and PoX such as high efficiency, low computational overhead, and high through-
put. However, the application of DAG is not mature. More research is required to substantiate its
practical value.
Different consensus protocols have different advantages, disadvantages, performance, and secu-
rity levels. In this way, different consensus protocols are applicable to different scenarios. Consensus
protocol such as PoW has high security levels and so can be deployed in public blockchain systems.
In addition, consensus protocol such as PBFT, which can achieve a high throughput, can only
be deployed in the permissioned blockchain systems since the limitation of its participant nodes.
Therefore, we provide a thorough comparison between different consensus protocols.

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 19

Table 4. Comparison between consensus protocols

Name Type Throughput Scalability Finality Adversary Tolerance Advantage Disadvantage Vulnerability Application
Low Transaction Cost Communication Overhead
BFT Permissioned High Low Deterministic 33.3% Replicas 33% Attack Tendermint
Instant Block Finality Centralization
High Throughput Communication Overhead
PBFT Permissioned High Low Deterministic 33.3% Faulty Replicas 33% Attack Hyperledger
Instant Block Finality Centralization
Low Throughput
Free to Join
PoW Permissionless Low Low Probabilistic 50% Computing Power Waste Energy Selfish Mining Bitcoin
Adaptive Consensus
High Fork Rate
Energy Efficient Communication Overhead
PoS Permissionless Low Low Probabilistic 50% Stake Long Range Attack Peercoin
Rolling Committee Matthew Effect
PoC Permissionless Low Low Probabilistic 50% Space Energy Efficient Waste Disk Space Selfish Mining IPFS
Trust Requirement
PoA Permissionless Low High Probabilistic 50% of Online Stake Energy Efficient Single Point Failure Decred
Limited Application Scenarios
PoI Permissionless Low Low Probabilistic 50% Stake Less Chance of Hoarding Trust Requirement Single Point Failure NEM
PoB Permissionless Low Low Probabilistic 50% Coins Long-Term Incentive Low Confirmation Latency Denial-of-Speding Attack XCP
DAG Permissioned High High Probabilistic 33.3% Computing Power High Throughput Communication Overhead Sybil Attack IOTA
Energy Efficient
Ripple Permissionless High High Deterministic 20% Faulty Nodes Trust Requirement Single Point Failure Ripple
Fast Block Finality

The BFT series consensus protocols can only be deployed in the permissioned blockchain system
so that it cannot be directly applied to the public chain. However, the BFT series consensus protocols
can achieve a high throughput compared to some other public chain consensus protocols such as
the Proof-of-Work (PoW) protocol.
The PoX series consensus protocols can be applied to the public chain. However, it cannot meet
the commercial usage performance demand. Moreover, most PoX series consensus protocols are
probabilistic so that the fork problem exists. However, a deterministic consensus protocol can
alleviate the blockchain system from the fork problem.
As to the performance, even with the same consensus protocol, slightly changing some parameters
may result in different performance. Accordingly, we talk about the performance of some specific
consensus protocols. The classical Nakamoto consensus could provide a 7 tx/s processing rate. In
Ethereum, the speed is 15 tx/s. As to the PoS protocol, Ouroboros system’s throughput can reach
256 tx/s while Snow-White can reach 110 tx/s. Moreover, in some hybrid consensus protocols,
Algorand can provide a 90 tx/s transaction processing rate. The Hyperledger can reach 110 tx/s
speed.
Considering the security levels of these protocols, most protocols do not provide a formal
security analysis of their protocols. This is because most of the agreements are just engineering
implementations rather than theoretical constructs. Specifically, Ouroboros provides a formal
security analysis and proves their protocol is secure and can reach a δ -Nash Equilibrium.
These consensus mechanisms also have some limitations. For example, the usage of PoW, DAG,
and PoC is limited by the total amount of computing power/space in the system, which means
that a system without enough computation/storage resource is vulnerable to 51% attack. As to
the BFT and PBFT, they can provide a high-throughput service for a few nodes. However, when
the number of nodes increases, it is hard to maintain the transaction processing rate because of
the heavy communication overhead. For PoS consensus protocol, there are situations where rich
people are getting richer, and poor people are getting poorer. This is because rich people have a
greater chance of releasing a new block, which in turn will bring them more benefits. Therefore,
the PoS consensus protocol needs a sustainable incentive mechanism, which can alleviate the PoS
protocol from centralization.

4.3 Consensus Characteristic in IoT Blockchain

In Table 2, the majority of IoT-blockchain application use PoW as their consensus mechanism.
However, in considering the compatibility between the limitations of IoT devices and consensus
mechanisms, PoW and the above common consensus mechanisms, which are the backbone for
reaching the global consensus, may not be perfect. Because the IoT devices have limited computing
capability, energy, and storage capacity, at the same time, the proof of work in the decentralized

, Vol. 1, No. 1, Article . Publication date: November 2019.

20 L.Lao et al.

network must be done to achieve validity and consistency. A suitable consensus mechanism
for IoT blockchain should achieve a balance between energy consumption, system performance,
and security. We summarize the characteristics of IoT blockchain as high energy efficiency and
lightweight consensus mechanism.

4.3.1 High Energy Efficiency. IoT blockchain needs to execute a consensus process efficiently.
Considering the limitations in storage and computing capability of IoT devices, light clients or no
miners are adopted for IoT blockchain. IOTA adopts DAG consensus: Each node in the network
who wants to generate transactions must actively participate in consensus process by approving
two previous transactions. Moreover, there are no miners in IOTA, which is more energy efficient.
Ambrosus [Ambrosus 2019] uses IoT sensors to collect client supply chain data, which is then
authenticated by Proof-of-Authority (PoA) consensus protocol. This protocol confers authentication
capabilities to the most trusted and high-integrity master nodes.
Helium [Haleem et al. 2018] adopts Honey Badger BFT which offers a perfect balance between
compromise and results. First, there are no timing assumptions: Honey Badger BFT assumes that
messages in a network eventually get delivered, which is usually an issue in a network. Second,
there is censorship resistance, which means that players cannot pick and choose whom to support
so that they can always win. In mining terms, it implies that miners cannot look into transactions
before they are published. Third, it is a permission-less consensus, so that anyone can join the
network.

4.3.2 Lightweight Mechanism. Atonomi [Atonomi 2019] partially adopts Ethereum network, even
though its transaction processing might become costly or speed-prohibitive. Adopting Ethereum as
its network means that Atonomi can use the well-deployed features of Ethereum immediately. To
maintain a healthy ecosystem, Atonomi can evaluate migration or simultaneous use of alternative
blockchain technologies for these transactions.
IoT Chain (ITC) [Database 2018] combines several technologies, such as Verifiable Virtual Routing
Forwarding (VRF), Practical Byzantine Fault Tolerance (PBFT), Directed Acyclic Graph (DAG), and
Simplified Payment Verification (SPV), to build a light operating system based on the blockchain.
The data in IoT Chain are stored in different node. To provide a high transaction processing rate,
ITC employs a hybrid consensus mechanism, PBFT and DAG. Specifically, the PBFT works for the
main chain and DAG works for the side chains.
Waltonchain [WaltonChain 2018] aims to provide a blockchain-based complete data exchange
and absolute information transparency for IoT. It adopts a parent-child chain architecture. Its
consensus mechanism is called WPoC (Waltonchain Proof of Contribution), which combines PoW,
PoS, and PoL (Proof-of-Labor). The parent chain uses the combination of PoW and PoS while the
data exchange between parent and child chains are processed by master node using the PoL.
Blockcloud [Ming et al. 2018] proposes a service-centric blockchain architecture to empower IoT.
It also designs a new consensus protocol called proof-of-service to provide decentralized trust and
security service. Proof-of-service is a hybrid consensus protocol, which combines the advantages of
both permissioned- and permissionless- protocol. It relies on the permissionless protocol to achieve
a Byzantine agreement and uses the permissioned protocol to improve the performance.

5 TRAFFIC MODEL OF IOT BLOCKCHAIN

In this section, we summarize traffic models of existing distributed systems, and provide our analysis
and traffic models of IoT blockchains.

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 21

5.1 Current Traffic Models

Traffic modeling work has been carried out from the Internet [Wu et al. 2018; Zhang et al. 2015,
2012, 2013] to IoT networks [Al-Turjman 2018; Al-Turjman et al. 2017; Masek et al. 2016]. In the
existing work of blockchain traffic models, the majority are based on peer-to-peer (P2P) traffic
models since the message exchange, message validation, and transaction process of blockchain are
all handled by peer nodes as in a P2P network. We will first summarize the existing work on P2P
traffic modeling and then illustrate the related work of blockchain traffic modeling.
5.1.1 P2P Network Modeling. A peer-to-peer (P2P) network consists of connected devices that
share resources between each others without centralized management. BitTorrent is a popular
application based on the P2P network. Qiu et al. [Qiu and Srikant 2004] introduce a fluid model
to analyze the performance of the BitTorrent system. The modeling results give insight into how
network performance is affected by different parameters. Kazaa is another famous P2P application.
Gummadi et al. [Gummadi et al. 2003] analyze Kazaa P2P file-sharing traffic in order to investigate
the nature of file-sharing workloads. Hefeeda et al. [Hefeeda and Saleh 2008] carry out an eight-
months traffic measurement to study the cache-related P2P traffic characteristics. The results show
that the traffic pattern of a P2P network follows a Mandelbrot-Zipf distribution. The distribution
shows the probability p(i) of a user accessing a resource at a popular order i out of N accessible
resources is:
K
p(i) = , (1)
(i + q)α
where K = 1/( i=1 1/(i + q)α ), with plateau factor q and skewness factor α. q defines the degree of
ÍN
smoothing of the distribution head. The smoothed distribution head represents the most desired
resources. These objects have characteristic of immutable and large file size, and they are suitable
for storing in the cache. When q = 0, the probability will follow a Zipf-like distribution with the
skewness factor α.
In addition, Li et al. [Li et al. 2011b] establish a P2P network traffic recognition model based on
the multi-dimensional support vector machine (MSVM). The model uses the MSVM as a classifier
to identify P2P traffic and Non-P2P traffic, and conduct feature extraction, data processing, and
MSVM training.
5.1.2 Blockchain Related Traffic Modeling. With the enormous growth in the field of blockchain
technology, there are different communication models. They have different traffic models and vary
in terms of algorithm, processing power, time and storage approach. We categorize them into three
types: Gossip protocol, Kademlia algorithm and Direct Acyclic Graph (DAG).

Gossip protocol
Gossip is a widely used protocol designed for P2P communications in distributed systems
including a large portion of current blockchain systems. In the gossip protocol, every IoT device
always chooses one or several random targets in each round before sending messages (namely
gossips). Eventually, all IoT devices receive the gossip twice in a round of data transmission. This
ensures data integrity even when the message sent is corrupted (by comparing the two received
messages). Therefore, the gossip protocol can serve as a robust and reliable message delivery
mechanism for general IoT systems.
However, the performance of gossip protocol suffers from the large message overhead in the
communication process. Fraigniaud et al. [Fraigniaud and Giakkoupis 2010] infer that when message
size is ρ, the time complexity of gossip protocol is O(loдn) and the message bit complexity of it is
Ω(n(ρ + loд(loдn)). Jenkins et al. [Jenkins et al. 2001] show that gossip overhead rises considerably

, Vol. 1, No. 1, Article . Publication date: November 2019.

22 L.Lao et al.

as the number of nodes in the network grows. For the p th node Np with infectivity I (Np ) and
susceptibility S(Np ) in a subgroup д, the expected number of messages sent from Np to the subgroup
д in a round is: Õ Õ
I (Np )( S(Np )). (2)
Np ∈д Nq
And the expected numbers of messages received by Np from the subgroup д in a round is
Õ Õ
S(Np )( I (Np )). (3)
Np ∈д Nq
Kademlia algorithm In Kademlia, every node is assigned a random and unique ID before joining
the network. Each ID consists of only either 1 or 0 and has a fixed length. Kademlia uses IDs
to calculates the logical distances between any two IoT devices by applying exclusive or (XOR)
operation on node IDs. Every node owns a random neighbor list, in which the IDs of nodes are
grouped into K-buckets based on the common prefixes. The K-buckets only stores the active nodes
or records. Compared with other methods (such as gossip), the approach used by Kademlia largely
reduces the computational overhead. For instance, when one node wishes to communicate with
other nodes, the message will only be relayed no more than n times to reach the target node in a
network with 2n nodes. Maymounkov et al. [Maymounkov and Mazieres 2002] summarize that
Kademlia only requires to exchange information with O(loдn) nodes during a search in a network
of n nodes. Cai et al. [Cai and Devroye 2013] give a mathematical proof to support this conclusion.
Consider a Kademlia network of n nodes with IDs from x 1, ..., x n and all IDs have the same length
d. For a node x ∈ x 1, ..., x n , let D i (x) be the set of nodes that share the same prefix with x with
prefix length d − i. The i t h bucket of this node contains k IDs randomly selected from D i (x). In
other words, filling i − th bucket of node x can be considered to be adding pointers from the leaf x
to k leaves chosen randomly from D i (x). Let Txy be the number of attempts needed by Kademlia to
search from leaf x to target ID y. Assume that x 1, ..., x n are chosen deterministically from {0, 1}d .
It can be proved that

log n
sup sup sup E[Txy ] ≤ (1 + o(1)) . (4)
x 1 ,...,x n x ∈ {x 1 ,...,x n } y ∈ {0,1}d Hk
In Eq.4, Hk is the k t h Harmonic Number. Since Hk /log k → ∞, when E[Txy ] is bounded by loдk n.
This justifies that in Kademlia only O(loдn) nodes are needed to contact with to search for a target
node.
The experiment on the performance of Kademlia is shown in [Brunner and Biersack 2006]. The
result shows that there is 60% probability that the Kademlia system can search and find the target
data within 15 seconds after a user publishes new data. Also, there is 100% probability that the
system can obtain the target data within 45 seconds. Another experiment evaluation of Kademlia is
shown in [Ou et al. 2010]. The result shows the Kademlia system is resilient against churns and
data corruption. Also, the paper concludes that the Kademlia system’s power consumption is low
with decent efficiency.

Direct Acyclic Graph (DAG) Directed Acyclic Graph (DAG) is a technique used in distributed
ledger technology. On a DAG network, the transactions are verified on a P2P basis. Each node
needs to verify the transactions of at least two previous nodes in order to perform a transaction.
And the ordered sequence of the verification events is called DirectedGraph.
Compared with the gossip protocol, DAG has no ring structure and it is unidirectional. It means
the performance of resource locating in DAG is guaranteed with a deterministic searching path.

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 23

block generation block propagation

block
verification
Node A

inv getdata block

or transaction

Node B

Fig. 6. Block generation and block propagation in blockchain.

DAG achieves higher search efficiency and computation compared with other communication
protocols that contain ring structure or non-directional structures.
DAG is capable of connecting thousands of IoT devices in an IoT system. Each device in
the system will be assigned a unique label for rapid positioning and communication. Combi-
natorial enumeration [Robinson 1977] illustrates the DAG labeling issue. The numbers on n
labeled nodes, for n = 0, 1, 2, 3, ... from the On-Line Encyclopedia of Integer Sequences are
1, 1, 3, 25, 543, 29281, 3781503, 1138779265, ...
Tangle [Tangle IOTA 2018] is an implementation of DAG. In Tangle, IoT devices are connected
with one another with a specified order and direction. When a device wishes to contact another
device, it can send a message directly to the target host. All devices in the network will be simulta-
neously updated. Tangle achieve a much higher overall throughput by allowing different branches
of the DAG to merge. Popov et al. [Popov 2016] analyze the mathematical foundations of IoTA,
Tangle and DAG. The paper infers that the total number of unapproved transactions in Tangle L in
a certain period of time can be calculated as

L 0 = 2λh. (5)
In the equation, λ is the expected event rate of Poisson distribution, and h is the average time
that a device needs to issue a transaction.
Benčić et al. [Benčić and Žarko 2018] compare DAG with traditional blockchain approaches.
They show that DAG has higher efficiency and scalability than traditional blockchain approaches.

5.2 Proposed Traffic Models

5.2.1 Blockchain Traffic Models. The traffic of blockchain can be categorized mainly into two
sections, block generation, and block propagation. Fig. 6 shows the message flow of block generation
and block propagation with the block generation process taking place before Node A receives a
block. After Node A receives a block, it verifies the block and sends it to neighbors. The block
propagation process takes place in turn.
In the block generation, based on Proof-of-Work [Decker and Wattenhofer 2013], a node needs
to calculate a value called nonce to publish a block. The nonce is a hash value included in the block
header. The node needs to try a large number of times to find a valid nonce. If the current target of
the nonce is 400 billion, a node will have 400 billion/2256 chances to solve the nonce and publish a
block. Blocks are generated around the world through the global computing power of nodes. The
difficulty of finding a valid nonce is dynamically adjusted according to the latest block generation
interval to keep the block generation interval remains relatively unchanged. On average, in bitcoin
network, a block is produced in every 10 minutes [Beccuti et al. 2017]. In Ethereum, the average

, Vol. 1, No. 1, Article . Publication date: November 2019.

24 L.Lao et al.

block produce time is 15 seconds. The difficulty of the mathematical problem increases over time
as the global computing power of miners increase. The expected time between blocks remains 10
minutes. It shows that the probability of generating a block satisfies the Poisson distribution. Let
λb be the success rate of generating a block, tc be a certain time interval, and N be the generated
blocks. The probability of generating N blocks in time tc follows a Poisson distribution:

(λb tc )N −λb tc
∞
∫
Pд = e dtc . (6)
0 N!
In block propagation, blocks and transactions are continuously propagated in the blockchain
network for updates and synchronization. To avoid message duplication, an inv message is sent
first [Decker and Wattenhofer 2013]. If a node receives an inv message that contains a not received
block hash or transaction hash in its inventory, it will respond with a getdata message. The sender
of inv message will transfer the block or transaction to the node after getting the getdata message.
On average, a block contains 500 transactions and the size of it is 1MB in Bitcoin [Bowden et al.
2018]. The total size of the inv and the getdata message is 120B on average. It is negligible when
compared to the block size. The blockchain transaction rate can be found on the official blockchain
website of Bitcoin. It shows there are 2.73 Bitcoin transactions added to the mempool per second.
Let Aд be the average arrival rate of blocks, Lb be the average size of blocks, Rb be the transmission
rate of a block, and Tp be the time period. The traffic intensity of block propagation on a single
node can be defined as:
Aд Lb
Tp . (7)
Rb
In addition, we also model the traffic distribution of propagating a new block to the entire
blockchain network. In different blockchain implementations, the traffic distribution differs with
variances in the communication protocols. In this paper, we mainly discuss the traffic distribution
of the propagation process in Bitcoins or Bitcoin-like blockchain systems. The gossip protocol is
considered when modeling propagation traffic.
Assume in the initial state of blockchain network with x + y nodes, a new message is generated
and acknowledged by y nodes. After that, the message should be propagated to the remaining x
nodes. A node will only receive the same message once. When a node receives a new message,
it will actively propagate the message to b neighbors. We assume every package can be sent and
received without error. And we assume the number of nodes remains stable for the entire process.
Initially, we have y = 1 and x = n. So the total number of nodes can also be presented as 1 + n.
During the process of a message being broadcast, x will decrease while the message y will increase.
Based on the gossip protocol, the decreasing rate of x should be dT dx
= −βxy. β is the contact rate
of all the (x, y) pair combination, which can be calculated by β = b/n. While x + y = n + 1, base on
the Epidemic Information Dissemination [Eugster et al. 2004] we can further obtain:

(n + 1)
y= . (8)
1 + ne −β (n+1)T
We define a constant c as c = b(n+1)
2n
. When T = c ∗loд(n), the number of nodes that have received
the new message can be estimated as y ≈ (n + 1) − n cb−2 1 1
. Since n cb−2 is relatively small, we can
assume that the propagation process ends at the time T = cloд(n). Since every node is broadcasting
the same message to its peers, the sizes of the received messages and sent messages are the same
as Smsд . During the propagation period, we calculate the traffic due to broadcasting blocks or
transactions, since the control messages are trivial compared to the total traffic. The traffic at time
T is corresponding to the number of received nodes. It can be calculated as:

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 25

(n + 1)Smsд
. T RFC(T ) = (9)
1 + ne −β (n+1)T
We simulate the infected nodes over time for the gossip protocol in Fig. 7. The simulation shows
a network with N nodes, and each node has h neighbors in its neighbor list. Each node will infect b
neighbors randomly in a gossip round. The experiment is conducted with inputs with b = 5, h = 10
and N = 2000, 5000, 10000. In Fig. 7, the horizontal axis shows the number of gossip rounds. We can
see that after seven gossip rounds, the first infected node can update 2000 nodes in the network.
The traffic can be calculated by the number of infected nodes times the average size of a packet in
a specific period of time.

N=2000 ,h=10,b=5
100% N=5000 ,h=10,b=5
N=10000,h=10,b=5

80%
Percentage of Infected Nodes

60%

40%

20%

0%
0 2 4 6 8
Gossip Rounds

Fig. 7. Percentage of infected nodes versus gossip rounds.

5.2.2 IoT-Blockchain Traffic Model. We divide the traffic model of IoT blockchain into two cate-
gories. They are different in the deployments of IoT devices, blockchain nodes, and their connections.
This refers to the physical layer, blockchain layer, and network layer of the 5-layer architecture for
IoT blockchain we proposed in section 3.4.
IoT Devices in the Blockchain Network Structure
In this structure, the IoT devices act as normal nodes in the blockchain network, as shown in Fig.
8. In the blockchain network, the IoT device nodes are considered as light nodes. They don’t need
to store a copy of the full blockchain and conduct the mining work.
This structure achieves maximum decentralization, since there is no single central node to
manage or monitor the nodes. Overall performance will not be adversely affected even if some
IoT nodes are corrupted. Also, there are pre-configured high-performance nodes are acting as
the backbone of the blockchain network. Work that requires heavy computation or massive data
storage will be assigned as the backbone nodes. Therefore, malfunctioning of the IoT device nodes
will not affect blockchain performance.
The rate of generating a block of traffic in this structure will follow Eq. 6. The traffic of nodes in
this structure can be calculated by Eq. 9.

, Vol. 1, No. 1, Article . Publication date: November 2019.

26 L.Lao et al.

IoT Device Nodes

Blockchain Network

Fig. 8. Structure of IoT devices in a blockchain network.

Central Proxy Server

IoT Device Nodes Blockchain Network

Fig. 9. Structure of IoT devices outside of a blockchain network.

IoT Devices Outside of the Blockchain Network Structure

In this structure, the blockchain network is considered as a separated network. A central base
station is placed to interact with the IoT device nodes directly. The requests sent by IoT devices
nodes will be processed by the base station and delivered to the blockchain network. When the
blockchain network confirmed the transaction, the IoT device will be notified. This structure is
illustrated in Fig. 9.
The benefit of this structure comes from its convenient configuration and maintenance. Since
the IoT device nodes are not directly connected to the blockchain network, the proxy server can
act as a traffic regulator. It becomes easier to control traffic and provide efficient transactions by
caching and reducing the possibility of data deluge. Compared to the structure of IoT devices in
a blockchain network, this structure allows the IoT system to utilize the functions of blockchain
without maintenance costs and configuration overhead of blockchain. The traffic model of the IoT
network depends on the types of IoT services. For the blockchain network, the traffic will follow
Eq. 6 and Eq. 9.

6 CONCLUSIONS
Blockchain technology has been extensively applied in diverse services. The Internet of Things
(IoT) is getting ubiquitous in modern life. Blockchain has provided a practical solution to address
many of the limitations of traditional IoT applications. The integration of the IoT and blockchain
creates a verifiable and secure network. This paper presents a comprehensive overview of IoT
blockchains, including typical architectures, communication protocols, applications, and traffic
models. This paper provides a good summary for researchers or practitioners who are interested in
understanding the concepts, functions and traffic models of IoT blockchain. Furthermore, we make

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 27

comparisons between current consensus algorithms and communication protocols and show their
strengths and shortcomings when applied in IoT blockchains. Finally, we analyze the current traffic
models of the P2P system and blockchain system. We also give traffic models of IoT blockchain
systems.
Considering that the number of IoT-blockchain applications grows each year, we believe that new
consensus mechanisms for enhancing the performance of IoT devices in the blockchain network
will be good directions in the coming year. Also, the solutions for solving scalability, processing
power, or storage problem of IoT device in the blockchain network is an interesting topic. Another
interesting future research direction is IoT-blockchain traffic model analysis. However, it may
require a large scale of simulation, data collection, and data analysis.
Apart from this, how to ensure security when integrating IoT with blockchain is still a challenging
topic because the IoT components do not have enough resources to perform the encryption
algorithms in the blockchain. Also, IoT devices are resource restricted to validate transitions and
maintain the correctness of the blockchain. Moreover, the access control of keeping the system
security while allowing numerous IoT components to join a private network is hard to achieve.
Additionally, regulation and development policies on the IoT-blockchain applications are still
missing. There is no instruction or rules to follow. Comprehensive regulation is needed to clarify
the rules and guide the integration of IoT systems and blockchain technology.

7 ACKNOWLEDGMENTS
This work was partially supported by the HK PolyU grant P0011609 and HK GRF PolyU 152124/19E.

REFERENCES
Ala Al-Fuqaha, Mohsen Guizani, Mehdi Mohammadi, Mohammed Aledhari, and Moussa Ayyash. 2015. Internet of things: A
survey on enabling technologies, protocols, and applications. IEEE Communications Surveys & Tutorials 17, 4 (2015),
2347–2376.
Fadi Al-Turjman. 2018. Information-centric framework for the Internet of Things (IoT): Traffic modeling & optimization.
Future Generation Computer Systems 80 (2018), 63–75.
Fadi Al-Turjman, Enver Ever, and Hadi Zahmatkesh. 2017. Green Femtocells in the IoT Era: Traffic Modeling and Challenges–
An Overview. IEEE Network 31, 6 (2017), 48–55.
Olivier Alphand, Michele Amoretti, Timothy Claeys, Simone Dall’Asta, Andrzej Duda, Gianluigi Ferrari, Franck Rousseau,
Bernard Tourancheau, Luca Veltri, and Francesco Zanichelli. 2018. IoTChain: A blockchain security architecture for the
Internet of Things. In Wireless Communications and Networking Conference (WCNC), 2018 IEEE. IEEE, Barcelona, Spain,
1–6.
Ambrosus 2019. Ambrosus - Enabling Sensors to Talk to Blockchain. https://ambrosus.com/.
Jeffrey G Andrews, Stefano Buzzi, Wan Choi, Stephen V Hanly, Angel Lozano, Anthony CK Soong, and Jianzhong Charlie
Zhang. 2014. What will 5G be? IEEE Journal on selected areas in communications 32, 6 (2014), 1065–1082.
Elli Androulaki, Artem Barger, Vita Bortnikov, Christian Cachin, Konstantinos Christidis, Angelo De Caro, David Enyeart,
Christopher Ferris, Gennady Laventman, Yacov Manevich, et al. 2018. Hyperledger fabric: a distributed operating system
for permissioned blockchains. In Proceedings of the Thirteenth EuroSys Conference. ACM, New York, NY, 30.
Andreas M Antonopoulos. 2014. Mastering Bitcoin: unlocking digital cryptocurrencies. " O’Reilly Media, Inc.", California, CA.
Giuseppe Ateniese, Michael T Chiaramonte, David Treat, Bernardo Magri, and Daniele Venturi. 2018. Hybrid blockchain.
US Patent 9,959,065.
Hany F Atlam, Ahmed Alenezi, Madini O Alassafi, and Gary Wills. 2018. Blockchain with internet of things: Benefits,
challenges, and future directions. International Journal of Intelligent Systems and Applications 10, 6 (2018), 40–48.
Atonomi 2019. Atonomi - Bringing Trust and Security to IoT. Retrieved March 20, 2019 from https://atonomi.io/
Attores 2019. Attores - Smart Contracts as a Service | Blockchain Singapore. Retrieved March 20, 2019 from https:
//attores.com/
Luigi Atzori, Antonio Iera, and Giacomo Morabito. 2010. The internet of things: A survey. Computer networks 54, 15 (2010),
2787–2805.
H Azima. 2018. NXT White Paper; New Transducer Ltd. Huntingdon PE18 6ED Harris, N (2018).
Arshdeep Bahga and Vijay K Madisetti. 2016. Blockchain platform for industrial internet of things. Journal of Software
Engineering and Applications 9, 10 (2016), 533.

, Vol. 1, No. 1, Article . Publication date: November 2019.

28 L.Lao et al.

Arati Baliga. 2017. Understanding blockchain consensus models. In Persistent. Persistent Systems Ltd., India.
Juan Beccuti, Christian Jaag, et al. 2017. The bitcoin mining game: On the optimality of honesty in proof-of-work consensus
mechanism. Technical Report. Swiss Economics.
Eli Ben-Sasson, Alessandro Chiesa, Daniel Genkin, Eran Tromer, and Madars Virza. 2013. SNARKs for C: Verifying program
executions succinctly and in zero knowledge. In Annual Cryptology Conference. Springer, Berlin, Heidelberg, 90–108.
Federico Matteo Benčić and Ivana Podnar Žarko. 2018. Distributed ledger technology: Blockchain compared to directed
acyclic graph. In 2018 IEEE 38th International Conference on Distributed Computing Systems (ICDCS). IEEE, Vienna, Austria,
1569–1570.
Dan Bieler. 2018. Blockchain and the Internet of Things: the IoT Blockchain Opportunity and Challenge. Retrieved February
27, 2019 from https://www.i-scoop.eu/blockchain-distributed-ledger-technology/blockchain-iot/
Carsten Bormann, Mehmet Ersue, and Ari Keranen. 2014. Terminology for constrained-node networks. Internet Engineering
Task Force (IETF): Fremont, CA, USA (2014), 2070–1721.
Rory Bowden, Holger Paul Keeler, Anthony E Krzesinski, and Peter G Taylor. 2018. Block arrivals in the Bitcoin blockchain.
arXiv preprint arXiv:1801.07447 (2018).
Richard Gendal Brown, James Carlyle, Ian Grigg, and Mike Hearn. 2016. Corda: an introduction. R3 CEV, August 1 (2016),
15.
René Brunner and E Biersack. 2006. A performance evaluation of the Kad-protocol. Institut Eurecom, France (2006).
Vitalik Buterin et al. 2014. A next-generation smart contract and decentralized application platform. white paper (2014).
Christian Cachin. 2016. Architecture of the hyperledger blockchain fabric. In Workshop on Distributed Cryptocurrencies and
Consensus Ledgers, Vol. 310. IBM Research, Zurich.
Xing Shi Cai and Luc Devroye. 2013. A probabilistic analysis of Kademlia networks. In International Symposium on Algorithms
and Computation. Springer, Berlin, Heidelberg, 711–721.
Miguel Castro, Barbara Liskov, et al. 1999. Practical Byzantine fault tolerance. In OSDI, Vol. 99. ACM, New Orleans, USA,
173–186.
Nutthakorn Chalaemwongwan and Werasak Kurutach. 2018. State of the art and challenges facing consensus protocols on
blockchain. In 2018 International Conference on Information Networking (ICOIN). IEEE, Chiang Mai, Thailand, 957–962.
Moumena A Chaqfeh and Nader Mohamed. 2012. Challenges in middleware solutions for the internet of things. In
Collaboration Technologies and Systems (CTS), 2012 International Conference on. IEEE, Denver, CO, USA, 21–26.
Cello Chen. 2018. AlipayHK and GCash Launch Cross-Border Remittance Service Powered by Alipay’s Blockchain
Technology. Retrieved February 20, 2019 from https://www.businesswire.com/news/home/20180625005561/en/
AlipayHK-GCash-Launch-Cross-Border-Remittance-Service-Powered
Deyan Chen and Hong Zhao. 2012. Data security and privacy protection issues in cloud computing. In 2012 International
Conference on Computer Science and Electronics Engineering, Vol. 1. IEEE, Hangzhou, China, 647–651.
Konstantinos Christidis and Michael Devetsikiotis. 2016. Blockchains and smart contracts for the internet of things. Ieee
Access 4 (2016), 2292–2303.
Anton Churyumov. 2016. Byteball: A Decentralized System for Storage and Transfer of Value. https://byteball.org/Byteball.
pdf.
Cisco IoT 2019. Internet of Things Cisco IoT is the bridge to business outcomes. Retrieved March 20, 2019 from
https://www.cisco.com/c/en/us/solutions/internet-of-things/overview.html
Marco Conoscenti, Antonio Vetro, and Juan Carlos De Martin. 2016. Blockchain for the Internet of Things: A systematic
literature review. In Computer Systems and Applications (AICCSA), 2016 IEEE/ACS 13th International Conference of. IEEE,
Agadir, Morocco, 1–6.
Hyperchain Corp. 2019. hyperchain white paper. https://hyperchain.readthedocs.io/zhCN/latest/consensus.html.
James Cowling, Daniel Myers, Barbara Liskov, Rodrigo Rodrigues, and Liuba Shrira. 2006. HQ replication: A hybrid
quorum protocol for Byzantine fault tolerance. In Proceedings of the 7th symposium on Operating systems design and
implementation. USENIX Association, Berkeley, CA, 177–190.
Michael Crosby, Pradan Pattanayak, Sanjeev Verma, Vignesh Kalyanaraman, et al. 2016. Blockchain technology: Beyond
bitcoin. Applied Innovation 2, 6-10 (2016), 71.
Li Da Xu, Wu He, and Shancang Li. 2014. Internet of things in industries: A survey. IEEE Transactions on industrial informatics
10, 4 (2014), 2233–2243.
Whitepaper Database. 2018. ITC White Paper Complete Version. https://whitepaperdatabase.com/iot-chain-itc-whitepaper/.
(2018).
Christian Decker and Roger Wattenhofer. 2013. Information propagation in the bitcoin network. In Peer-to-Peer Computing
(P2P), 2013 IEEE Thirteenth International Conference on. IEEE, Trento, Italy, 1–10.
DokChain 2019. DokChain | PokitDok. Retrieved March 12, 2019 from https://pokitdok.com/dokchain/
Ali Dorri, Salil S Kanhere, and Raja Jurdak. 2016. Blockchain in internet of things: challenges and solutions. arXiv preprint
arXiv:1608.05187 (2016).

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 29

Ali Dorri, Salil S Kanhere, Raja Jurdak, and Praveen Gauravaram. 2017a. Blockchain for IoT security and privacy: The
case study of a smart home. In Pervasive Computing and Communications Workshops (PerCom Workshops), 2017 IEEE
International Conference on. IEEE, Kona, HI, USA, 618–623.
Ali Dorri, Salil S Kanhere, Raja Jurdak, and Praveen Gauravaram. 2017b. LSB: A Lightweight Scalable BlockChain for IoT
Security and Privacy. arXiv preprint arXiv:1712.02969 (2017).
Kevin Driscoll, Brendan Hall, Håkan Sivencrona, and Phil Zumsteg. 2003. Byzantine fault tolerance, from theory to reality.
In International Conference on Computer Safety, Reliability, and Security. Springer, Berlin, Heidelberg, 235–248.
Cynthia Dwork and Moni Naor. 1992. Pricing via processing or combatting junk mail. In Annual International Cryptology
Conference. Springer, Berlin, Heidelberg, 139–147.
Stefan Dziembowski, Sebastian Faust, Vladimir Kolmogorov, and Krzysztof Pietrzak. 2015. Proofs of Space. In Advances in
Cryptology – CRYPTO 2015. Springer, Berlin, Heidelberg, 585–605.
Ariel Ekblaw, Asaph Azaria, John D Halamka, and Andrew Lippman. 2016. A Case Study for Blockchain in Health-
care:‘MedRec’ prototype for electronic health records and medical research data. In Proceedings of IEEE open & big data
conference, Vol. 13. IEEE, WA, USA, 13.
ElectriCChain 2018. ElectriCChain The Solar Energy Blockchain Project for Climate Change and Beyond. Retrieved March
20, 2019 from https://www.electricchain.org/
Patrick T Eugster, Rachid Guerraoui, A-M Kermarrec, and Laurent Massoulié. 2004. Epidemic information dissemination in
distributed systems. Computer 37, 5 (2004), 60–67.
Factom 2019. Factom | A blockchain innovations company. Retrieved March 20, 2019 from https://www.factom.com/
Filament. 2018. Filament’s Industrial Internet of Things Blockchain Solution Wins 2018
IoT Innovator Award. https://globenewswire.com/news-release/2018/09/26/1576581/0/en/
Filament-s-Industrial-Internet-of-Things-Blockchain-Solution-Wins-2018-IoT-Innovator-Award.html.
Pierre Fraigniaud and George Giakkoupis. 2010. On the bit communication complexity of randomized rumor spreading. In
Proceedings of the twenty-second annual ACM symposium on Parallelism in algorithms and architectures. ACM, New York,
NY, 134–143.
Jake Frankenfield. 2018. Proof of Burn. https://www.investopedia.com/terms/p/proof-burn-cryptocurrency.
S. Gao, Z. Li, Z. Peng, and B. Xiao. 2019. Power Adjusting and Bribery Racing: Novel Mining Attacks in the Bitcoin System.
In Proceedings of the 2019 ACM SIGSAC Conference on Computer and Communications Security (CCS ’19). ACM, New York,
NY, 833–850.
Arthur Gervais, Ghassan O Karame, Vedran Capkun, and Srdjan Capkun. 2014. Is bitcoin a decentralized currency? IEEE
security & privacy 12, 3 (2014), 54–60.
Jayavardhana Gubbi, Rajkumar Buyya, Slaven Marusic, and Marimuthu Palaniswami. 2013. Internet of Things (IoT): A
vision, architectural elements, and future directions. Future generation computer systems 29, 7 (2013), 1645–1660.
Krishna P Gummadi, Richard J Dunn, Stefan Saroiu, Steven D Gribble, Henry M Levy, and John Zahorjan. 2003. Measurement,
modeling, and analysis of a peer-to-peer file-sharing workload. ACM SIGOPS Operating Systems Review 37, 5 (2003),
314–329.
Amir Haleem, Andrew Allen, Andrew Thompson, Marc Nijdam, and Rahul Garg. 2018. A Decentralized Machine Network.
(2018).
Mohamed Hefeeda and Osama Saleh. 2008. Traffic modeling and proportional partial caching for peer-to-peer systems.
IEEE/ACM Transactions on networking 16, 6 (2008), 1447–1460.
Y. Huang, J. Zhang, J. Duan, B. Xiao, F. Ye, and Y. Yang. 2019. Resource Allocation and Consensus on Edge Blockchain in
Pervasive Edge Computing Environments. In 2019 IEEE 39Th international conference on distributed computing systems.
IEEE.
HUAWEI IOT 2019. Enterprise IoT Leading IoT, Driving Industry Digital Transformation. Retrieved March 20, 2019 from
https://e.huawei.com/en/solutions/technical/iot
Steve Huckle, Rituparna Bhattacharya, Martin White, and Natalia Beloff. 2016. Internet of things, blockchain and shared
economy applications. Procedia computer science 98 (2016), 461–466.
Seyoung Huh, Sangrae Cho, and Soohyung Kim. 2017. Managing IoT devices using blockchain platform. In Advanced
Communication Technology (ICACT), 2017 19th International Conference on. IEEE, Bongpyeong, South Korea, 464–467.
IOTA 2019. The Next Generation of Distributed Ledger Technology | IOTA. Retrieved March 12, 2019 from https:
//www.iota.org/
IPFS 2019. IPFS powers the Distributed Web. https://ipfs.io/.
Reuben Jackson. 2018. Why IoT Needs the Blockchain, and Blockchain needs IoT.
JD. 2018. The JD. http://ledger.jd.com/.
JDChain 2019. JD enterprise blockchain service. Retrieved March 12, 2019 from http://blockchain.jd.com/blockchain_store/
pc/index.html#/BlockChainTrace

, Vol. 1, No. 1, Article . Publication date: November 2019.

30 L.Lao et al.

Kate Jenkins, Ken Hopkinson, and Ken Birman. 2001. A gossip protocol for subgroup multicast. In Distributed Computing
Systems Workshop, 2001 International Conference on. IEEE, Mesa, AZ, USA, USA, 25–30.
Jiaojiao Jiang, Sheng Wen, Shui Yu, Yang Xiang, and Wanlei Zhou. 2016. Identifying propagation sources in networks:
State-of-the-art and comparative studies. IEEE Communications Surveys & Tutorials 19, 1 (2016), 465–481.
Zura Kakushadze and Ronald P Russo Jr. 2018. Blockchain: Data Malls, Coin Economies and Keyless Payments. (2018).
Jiawen Kang, Rong Yu, Xumin Huang, Sabita Maharjan, Yan Zhang, and Ekram Hossain. 2017. Enabling localized peer-
to-peer electricity trading among plug-in hybrid electric vehicles using consortium blockchains. IEEE Transactions on
Industrial Informatics 13, 6 (2017), 3154–3164.
Minhaj Ahmad Khan and Khaled Salah. 2018. IoT security: Review, blockchain solutions, and open challenges. Future
Generation Computer Systems 82 (2018), 395–411.
Rafiullah Khan, Sarmad Ullah Khan, Rifaqat Zaheer, and Shahid Khan. 2012. Future internet: the internet of things
architecture, possible applications and key challenges. In Frontiers of Information Technology (FIT), 2012 10th International
Conference on. IEEE, Islamabad, India, 257–260.
Aggelos Kiayias, Alexander Russell, Bernardo David, and Roman Oliynykov. 2017. Ouroboros: A provably secure proof-of-
stake blockchain protocol. In Annual International Cryptology Conference. Springer, Cham, 357–388.
Sunny King and Scott Nadal. 2012. Ppcoin: Peer-to-peer crypto-currency with proof-of-stake. self-published paper, August
19 (2012).
Ramakrishna Kotla, Lorenzo Alvisi, Mike Dahlin, Allen Clement, and Edmund Wong. 2007. Zyzzyva: speculative byzantine
fault tolerance. ACM SIGOPS Operating Systems Review 41, 6 (2007), 45–58.
Tim Kozak. 2018. Consensus Protocols That Meet Different Business Demands. https://blockchain.intellectsoft.net/blog/
consensus-protocols-that-meet-different-business-demands/.
Nir Kshetri. 2017. Can blockchain strengthen the internet of things? IT Professional 19, 4 (2017), 68–72.
Leslie Lamport, Robert Shostak, and Marshall Pease. 1982. The Byzantine generals problem. ACM Transactions on
Programming Languages and Systems (TOPLAS) 4, 3 (1982), 382–401.
Daniel Larimer. 2014. Delegated proof-of-stake (dpos). Bitshare whitepaper (2014).
In Lee and Kyoochun Lee. 2015. The Internet of Things (IoT): Applications, investments, and challenges for enterprises.
Business Horizons 58, 4 (2015), 431–440.
LeewayHertz 2019. Blockchain Development for Startups and Enterprises | USA | UAE. Retrieved March 12, 2019 from
https://www.leewayhertz.com/
Sergio Demian Lerner. 2015. DagCoin: a cryptocurrency without blocks. https://bitslog.files.wordpress.com/2015/09/
dagcoin-v41.pdf.
Xu Li, Rongxing Lu, Xiaohui Liang, Xuemin Shen, Jiming Chen, and Xiaodong Lin. 2011a. Smart community: an internet of
things application. IEEE Communications Magazine 49, 11 (2011).
Xiaoyong Li, Feng Zhou, and Xudong Yang. 2011b. A multi-dimensional trust evaluation model for large-scale P2P computing.
J. Parallel and Distrib. Comput. 71, 6 (2011), 837–847.
Zhetao Li, Jiawen Kang, Rong Yu, Dongdong Ye, Qingyong Deng, and Yan Zhang. 2017. Consortium blockchain for secure
energy trading in industrial internet of things. IEEE transactions on industrial informatics 14, 8 (2017), 3690–3700.
Iuon-Chang Lin and Tzu-Chun Liao. 2017. A Survey of Blockchain Security Issues and Challenges. IJ Network Security 19, 5
(2017), 653–659.
Erik Linask. 2018. Blockchain Momentum Continues, Will Reach $12 Billion in 2020. In The Blockchain Domain.
Bin Liu, Xiao Liang Yu, Shiping Chen, Xiwei Xu, and Liming Zhu. 2017. Blockchain based data integrity service framework
for IoT data. In Web Services (ICWS), 2017 IEEE International Conference on. IEEE, Honolulu, HI, USA, 468–475.
Liu Liu, Olivier De Vel, Qing-Long Han, Jun Zhang, and Yang Xiang. 2018. Detecting and preventing cyber insider threats: a
survey. IEEE Communications Surveys & Tutorials 20, 2 (2018), 1397–1417.
LO3 2018. LO3 Energy The Future of Energy. Retrieved March 12, 2019 from https://lo3energy.com/
Pavel Masek, Jan Masek, Petr Frantik, Radek Fujdiak, Aleksandr Ometov, Jiri Hosek, Sergey Andreev, Petr Mlynek, and
Jiri Misurec. 2016. A harmonized perspective on transportation management in smart cities: The novel IoT-driven
environment for road traffic modeling. Sensors 16, 11 (2016), 1872.
Petar Maymounkov and David Mazieres. 2002. Kademlia: A peer-to-peer information system based on the xor metric. In
International Workshop on Peer-to-Peer Systems. Springer, Berlin, Heidelberg, 53–65.
Dejan S Milojicic, Vana Kalogeraki, Rajan Lukose, Kiran Nagaraja, Jim Pruyne, Bruno Richard, Sami Rollins, and Zhichen
Xu. 2002. Peer-to-peer computing.
Zhongxing Ming, Shu Yang, Qi Li, Dan Wang, Mingwei Xu, Ke Xu, and Laizhong Cui. 2018. Blockcloud: A Blockchain-based
Service-centric Network Stack. https://www.block-cloud.io/blockcloudtechnicalwhitepaper.pdf. (2018).
Daniele Miorandi, Sabrina Sicari, Francesco De Pellegrini, and Imrich Chlamtac. 2012. Internet of things: Vision, applications
and research challenges. Ad hoc networks 10, 7 (2012), 1497–1516.

, Vol. 1, No. 1, Article . Publication date: November 2019.

A Survey of IoT Applications in Blockchain Systems 31

M.Padma, N.KasiViswanath, and T.Swathi. 2019. Blockchain for IoT Application: Challenges and Issues. International
Journal of Recent Technology and Engineering 7 (2019), 40–48.
MuleSoft IoT 2019. Solutions for IoT Extend connectivity from your enterprise and the cloud to devices at the edge of your
network. Retrieved March 20, 2019 from https://www.mulesoft.com/integration-solutions/api/iot
Satoshi Nakamoto. 2008. Bitcoin: A peer-to-peer electronic cash system. (2008).
Alicia Naumoff. 2016. Why Blockchain Needs ‘Proof of Authority’ Instead of ‘Proof of Stake’. https://cointelegraph.com/
news/why-blockchain-needs-proof-of-authority-instead-of-proof-of-stake.
NEM. 2018. NEM whitepaper. https://nem.io/wp-content/themes/nem/files/NEM_techRef.pdf.
Oracle IoT 2019. Accelerate Your Business with the Power of the Internet of Things. Retrieved March 20, 2019 from
https://www.oracle.com/solutions/internet-of-things/
Zhonghong Ou, Erkki Harjula, Otso Kassinen, and Mika Ylianttila. 2010. Performance evaluation of a Kademlia-based
communication-oriented P2P system under churn. Computer Networks 54, 5 (2010), 689–705.
Alfonso Panarello, Nachiket Tapas, Giovanni Merlino, Francesco Longo, and Antonio Puliafito. 2018. Blockchain and IoT
Integration: A Systematic Survey. Sensors (2018), 2575.
Zhe Peng, Haotian Wu, Bin Xiao, and Songtao Guo. 2019. VQL: Providing Query Efficiency and Data Authenticity in
Blockchain Systems. In 2019 IEEE 35th International Conference on Data Engineering Workshops (ICDEW). IEEE, Macao,
1–6.
Marc Pilkington. 2016. 11 Blockchain technology: principles and applications. Research handbook on digital transformations
(2016), 225.
Suporn Pongnumkul, Chaiyaphum Siripanpornchana, and Suttipong Thajchayapong. 2017. Performance analysis of private
blockchain platforms in varying workloads. In 2017 26th International Conference on Computer Communication and
Networks (ICCCN). IEEE, Vancouver, BC, Canada, 1–6.
Serguei Popov. 2016. The tangle. cit. on (2016), 131.
Giulio Prisco. 2016. Slock. it to Introduce Smart Locks Linked to Smart Ethereum Contracts, Decentralize the Sharing
Economy. Bitcoin Magazine. Nov-2015 [Online]. Available: https://bitcoinmagazine.com/articles/slock-it-to-introduce-smart-
locks-linked-to-smart-ethereum-contracts-decentralize-the-sharing-economy-1446746719. (2016).
Dongyu Qiu and Rayadurgam Srikant. 2004. Modeling and performance analysis of BitTorrent-like peer-to-peer networks.
34, 4 (2004), 367–378.
Gowri Sankar Ramachandran and Bhaskar Krishnamachari. 2018. Blockchain for the IoT: Opportunities and challenges.
arXiv preprint arXiv:1805.02818 (2018).
Ana Reyna, Cristian Martín, Jaime Chen, Enrique Soler, and Manuel Díaz. 2018. On blockchain and its integration with IoT.
Challenges and opportunities. Future Generation Computer Systems 88 (2018), 173–190.
Robert W Robinson. 1977. Counting unlabeled acyclic digraphs. In Combinatorial mathematics V. Springer, 28–43.
Gokhan Sagirlar, Barbara Carminati, Elena Ferrari, John D Sheehan, and Emanuele Ragnoli. 2018. Hybrid-IoT: Hybrid
Blockchain Architecture for Internet of Things-PoW Sub-blockchains. arXiv preprint arXiv:1804.03903 (2018).
Omar Said and Mehedi Masud. 2013. Towards internet of things: Survey and future vision. International Journal of Computer
Networks 5, 1 (2013), 1–17.
Mehrdad Salimitari and Mainak Chatterjee. 2018. An Overview of Blockchain and Consensus Protocols for IoT Networks.
arXiv preprint arXiv:1809.05613 (2018).
Mayra Samaniego and Ralph Deters. 2016. Blockchain as a Service for IoT. In Internet of Things (iThings) and IEEE Green
Computing and Communications (GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart
Data (SmartData), 2016 IEEE International Conference on. IEEE, Chengdu, China, 433–436.
Rüdiger Schollmeier. 2001. A definition of peer-to-peer networking for the classification of peer-to-peer architectures and
applications. In Peer-to-Peer Computing, 2001. Proceedings. First International Conference on. IEEE, Linkoping, Sweden,
Sweden, 101–102.
Ludwig Seitz, Goeran Selander, Erik Wahlstroem, Samuel Erdtman, and Hannes Tschofenig. 2017. Authentication and
authorization for constrained environments (ace). Internet Engineering Task Force, Internet-Draft draft-ietf-aceoauth-
authz-07 (2017).
Pradip Kumar Sharma, Saurabh Singh, Young-Sik Jeong, and Jong Hyuk Park. 2017. Distblocknet: A distributed blockchains-
based secure sdn architecture for iot networks. IEEE Communications Magazine 55, 9 (2017), 78–85.
Janusz J Sikorski, Joy Haughton, and Markus Kraft. 2017. Blockchain technology in the chemical industry: Machine-to-
machine electricity market. Applied Energy 195 (2017), 234–246.
Rashmi Sharan Sinha, Yiqiao Wei, and Seung-Hoon Hwang. 2017. A survey on LPWA technology: LoRa and NB-IoT. Ict
Express 3, 1 (2017), 14–21.
sloct.it 2018. slock.it A Blockchain Company. Retrieved March 20, 2019 from https://slock.it/
Biljana L Risteska Stojkoska and Kire V Trivodaliev. 2017. A review of Internet of Things for smart home: Challenges and
solutions. Journal of Cleaner Production 140 (2017), 1454–1464.

, Vol. 1, No. 1, Article . Publication date: November 2019.

32 L.Lao et al.

Nan Sun, Jun Zhang, Paul Rimba, Shang Gao, Leo Yu Zhang, and Yang Xiang. 2018. Data-driven cybersecurity incident
prediction: A survey. IEEE Communications Surveys & Tutorials 21, 2 (2018), 1744–1772.
Girish Suryanarayana and Richard N Taylor. 2004. A survey of trust management and resource discovery technologies in
peer-to-peer applications. (2004).
Melanie Swan. 2015. Blockchain: Blueprint for a new economy. "O’Reilly Media, Inc.".
Lu Tan and Neng Wang. 2010. Future internet: The internet of things. In Advanced Computer Theory and Engineering
(ICACTE), 2010 3rd International Conference on, Vol. 5. IEEE, Chengdu, China, V5–376.
Tangle IOTA 2018. Meet the Tangle. https://www.iota.org/research/meet-the-tangle.
Teachracers 2019. Blockchain Development Company | Blockchain Services and Solutions. Retrieved March 20, 2019 from
https://www.techracers.com/
TraceRx 2019. Tracerx : Global Blockchain Supply Chain for Drugs. Retrieved March 10, 2019 from https://www.leewayhertz.
com/project/tracerx/
UniquID 2018. UniquID Incorporation Blockchain Identity Access Management. Retrieved March 20, 2019 from https:
//uniquid.com/
Mališa Vučinić, Bernard Tourancheau, Franck Rousseau, Andrzej Duda, Laurent Damon, and Roberto Guizzetti. 2015.
OSCAR: Object security architecture for the Internet of Things. Ad Hoc Networks 32 (2015), 3–16.
Marko Vukolić. 2015. The quest for scalable blockchain fabric: Proof-of-work vs. BFT replication. In International Workshop
on Open Problems in Network Security. Springer, Cham, 112–125.
WaltonChain. 2018. WaltonChain White Paper V2.0. (2018).
Sheng Wen, Mohammad Sayad Haghighi, Chao Chen, Yang Xiang, Wanlei Zhou, and Weijia Jia. 2014. A sword with two
edges: Propagation studies on both positive and negative information in online social networks. IEEE Trans. Comput. 64,
3 (2014), 640–653.
Andrew Whitmore, Anurag Agarwal, and Li Da Xu. 2015. The Internet of Things-A survey of topics and trends. Information
Systems Frontiers 17, 2 (2015), 261–274.
Gavin Wood. 2014. Ethereum: A secure decentralised generalised transaction ledger. Ethereum project yellow paper 151
(2014), 1–32.
Miao Wu, Ting-Jie Lu, Fei-Yang Ling, Jing Sun, and Hui-Ying Du. 2010. Research on the architecture of Internet of Things.
In Advanced Computer Theory and Engineering (ICACTE), 2010 3rd International Conference on, Vol. 5. IEEE, Chengdu,
China, V5–484.
Tingmin Wu, Sheng Wen, Yang Xiang, and Wanlei Zhou. 2018. Twitter spam detection: Survey of new approaches and
comparative study. Computers & Security 76 (2018), 265–284.
Xage 2018. Home page of Xage Security. Retrieved March 20, 2019 from https://xage.com/
Jinhong Xie and Steven M Shugan. 2001. Electronic tickets, smart cards, and online prepayments: When and how to advance
sell. Marketing Science 20, 3 (2001), 219–243.
Zhihong Yang, Yingzhao Yue, Yu Yang, Yufeng Peng, Xiaobo Wang, and Wenji Liu. 2011. Study and application on the
architecture and key technologies for IOT. In Multimedia Technology (ICMT), 2011 International Conference on. IEEE,
Hangzhou, China, 747–751.
Andrea Zanella, Nicola Bui, Angelo Castellani, Lorenzo Vangelista, and Michele Zorzi. 2014. Internet of things for smart
cities. IEEE Internet of Things journal 1, 1 (2014), 22–32.
Jun Zhang, Xiao Chen, Yang Xiang, Wanlei Zhou, and Jie Wu. 2015. Robust network traffic classification. IEEE/ACM
Transactions on Networking (TON) 23, 4 (2015), 1257–1270.
Jun Zhang, Yang Xiang, Yu Wang, Wanlei Zhou, Yong Xiang, and Yong Guan. 2012. Network traffic classification using
correlation information. IEEE Transactions on Parallel and Distributed systems 24, 1 (2012), 104–117.
Jun Zhang, Yang Xiang, Yu Wang, Wanlei Zhou, Yong Xiang, and Yong Guan. 2013. Network traffic classification using
correlation information. IEEE Transactions on Parallel and Distributed Systems 24, 1 (2013), 104–117.
Yu Zhang and Jiangtao Wen. 2017. The IoT electric business model: Using blockchain technology for the internet of things.
Peer-to-Peer Networking and Applications 10, 4 (2017), 983–994.
Zibin Zheng, Shaoan Xie, Hongning Dai, Xiangping Chen, and Huaimin Wang. 2017. An overview of blockchain technology:
Architecture, consensus, and future trends. In Big Data (BigData Congress), 2017 IEEE International Congress on. IEEE,
Milan, ITALY, 557–564.
Zibin Zheng, Shaoan Xie, Hong-Ning Dai, Xiangping Chen, and Huaimin Wang. 2018. Blockchain challenges and opportu-
nities: A survey. International Journal of Web and Grid Services 14, 4 (2018), 352–375.

, Vol. 1, No. 1, Article . Publication date: November 2019.