Multi-Chain Architecture for Blockchain Networks

∗
Kristián Košt’ál
Institute of Computer Engineering and Applied Informatics
Faculty of Informatics and Information Technologies
Slovak University of Technology in Bratislava
Ilkovičova 2, 842 16 Bratislava, Slovakia
[email protected]

Abstract Keywords
This paper focuses on interoperability between blockchain blockchain, cross-chain, multi-chain, atomic swaps, inter-
networks, whether existing or new. The issue of interop- operability
erability is a very current topic and is the subject of re-
search by many articles or institutions. The current trend
in the articles is to address interoperability only between 1. Introduction
homogeneous networks, i.e. the so-called Sharding, and We live in a digital age. The Internet is used for everyday
not at all dealing with interoperability between different communication, bill payments, shopping, charity support
networks. The second trend is to focus on two specific or just for fun. Among them, blockchain technology cre-
blockchain networks and interconnect them as efficiently ates another digital dimension and offers new business
as possible. However, both of these trends are flawed, as opportunities.
a large number of new heterogeneous networks have re-
cently emerged, which are incompatible with each other, A key feature is storing data and transactions that change
so that all processes take place in isolation. If they want to it in block strings. It acts as a distributed database that
communicate with another network, this is almost impos- processes an ever-growing list of transactions for which
sible because there is no standard for interlock blockchain no authority owns the data. Thanks to this function, the
communication. The only option is to exchange assets blockchain is extremely resistant to unauthorized modifi-
within the exchange, but this is only a scenario for cryp- cations to the stored data. The financial sector was there-
tocurrencies and not for other use cases. Based on the fore the largest supporter. Publicly, the most perceived
research, an architecture is proposed that uses an API to blockchain applications are cryptocurrencies [8, 24]. Even
connect existing blockchain networks and allows commu- some of the most dominant regulators in the United States,
nication between any blockchain. From this blockchain it such as Commodity Futures Trading Commission (CFTC)
is possible to derive other blockchains that are scalable. and Securities and Exchange Commission (SEC), see the
The architecture has been implemented and experimen- positive potential of this technology.
tal verification has shown scalability up to 18,000 trans-
actions per second. In addition to its use in cryptocurrencies transactions,
the benefits of blockchains, such as decentralized infor-
Categories and Subject Descriptors mation verification and tamper resistance, have been no-
C.2.1 [Networks]: Network Architecture and Design— ticed by various sectors [20, 21, 10]. Important applica-
distributed networks; C.2.4 [Networks]: Distributed Sys- tions include the value register, the valuable website and
tems—distributed applications, distributed databases; value ecosystem services. Related application industries
D.2.12 [Software Engineering]: Interoperability; H.2.4 include logistics, financial systems, medical records, data
[Database Management]: Systems—distributed collection and verification in Internet of Things (IoT),
databases supply chain management, stock or option trading, so-
cial networking software, electronic patient records, mi-
cropayments / mobile payment systems , asset transac-
∗Recommended by thesis supervisor: Prof. Ivan Kotuliak
tions and distribution of digital products. People hope
Defended at Faculty of Informatics and Information Tech- that blockchains will be able to play the role of trusted
nologies, Slovak University of Technology in Bratislava on machines in the operation of these systems. Keeping a
September 22, 2020.
detailed record of related information and resolving in-
c Copyright 2020. All rights reserved. Permission to make digital formation asymmetry problems (where one party to an
or hard copies of part or all of this work for personal or classroom use economic transaction owns more related information than
is granted without fee provided that copies are not made or distributed
for profit or commercial advantage and that copies show this notice on the other party) will allow a reliable record to be created.
the first page or initial screen of a display along with the full citation. In the cases used above, a large amount of information
Copyrights for components of this work owned by others than ACM will need to be recorded on the blockchain [14].
must be honored. Abstracting with credit is permitted. To copy other-
wise, to republish, to post on servers, to redistribute to lists, or to use We will need more robust blockchain networks than are
any component of this work in other works requires prior specific per- used today to store large amounts of data. This robust-
mission and/or a fee. Permissions may be requested from STU Press,
Vazovova 5, 811 07 Bratislava, Slovakia. ness will mainly face problems with scalability, speed of
transactions and also privacy challenges [28, 25]. The
Košt’ál, K. Multi-Chain Architecture for Blockchain Networks. In-
formation Sciences and Technologies Bulletin of the ACM Slovakia, problem is also in the blockchain community, as there are
Vol. 12, No. 2 (2020) 8-14 many more initiatives aimed at creating new blockchains
 Information Sciences and Technologies Bulletin of the ACM Slovakia, Vol. 12, No. 2 (2020) 8-14 9

or cryptocurrencies than at developing possible connec- value. A technique where operations on two inter-
tions and means of communication between existing block- operating chains are set up in a way that they have
chain networks. For such a solution we need interoperabil- the same trigger which is usually a revelation of the
ity [11, 22] between today’s blockchains and that is the pre-image of a particular hash [3]. Hash-locking is
main focus of this work. simpler compared to relay schemes in which some
copy of one ledger must be stored on another ledger
because hash-locking only require an exchange of a
2. State of the Art
single hash between two ledgers thus requires signif-
Blockchain is, simply put, the technology of distributed icant less exchange of data to achieve interoperabil-
ledger book (DLT). Blockchain is a decentralized distribut- ity. The primary focus of hash-locking is to allow
ed database, which is stored as a chain of blocks, that atomic swaps between ledgers without requiring a
are cryptographically ordered. The blocks contain trans- third party as it is with notary schemes [16]. The
actions between addresses of users. All of this happens key drawback of hash-locking is that the underly-
in a peer-to-peer network. There exist more kinds of ing blockchain must support a special type of smart
blockchain technology: public, private, consortium. Fur- contracts, called Hash-TimeLock Contracts.
thermore, these can be divided by two groups: permis-
sionless, permissioned. When we talk about interoperabil-
ity, we think communication between two distinct block- These are three main techniques. Moreover, we can define
chains by any means. Therefore, we need to analyze the 4 potential use-cases for interoperability of blockchains as
options and techniques to achieve inter-operable ledgers. laid out in [3] and [16]:

2.1 Cross-chain
• asset portability - transfer asset from one block-
We can understand cross-chain technology as a bridge chain to another;
connecting more blockchains. Its main application is the
implementation of atomic cross-chain swaps between dif- • atomic swap - swap asset between two blockchains
ferent blockchains, asset conversion or inter-block com- in atomic way;
munication [4]. There are obvious obstacles to the dis-
tribution of value between blockchains. The cross-chain • cross-chain oracles - are used to read data from
is a complex distributed coordination task. It needs not outside of blockchain so it can be act upon them;
only a separate authentication capability for blockchain
• cross-chain asset encumbrance - lock up assets
nodes, but also decentralized input, as well as data ac-
within one ledger based on locking conditions de-
quisition and authentication beyond the blockchain [26].
pendent on another ledger.
Cross-chain technologies currently include [3]:

The techniques and use-cases are summarized in Table 1.

• Notary schemes - one side decides to take any
action on chain B when a specific event on chain 2.2 Related Work
A occurs. Means that a group of trusted parties
There exist more solutions which try to make interoper-
comes to Byzantine-Fault Toleracne (BFT) consen-
abilty between blockchains. The nowadays state of the art
sus over the state of the ledger according to some
solutions in this research area, according to my analysis
action on another ledger. The notaries usually exist
and best knowledge, are Cosmos [17] and Polkadot [27].
on separate ledger that they manage to ensure inter-
They are both trying to solve similar problems, concretely
operability between two or more ledgers [16]. This
introducing interconnection between any blockchain net-
is technologically one of the simplest solution to ac-
work with the help of centralized Cosmos’s Hub (or Polka-
complish interoperability goal under its trust model.
dot’s Relay Chain). To the best of my knowledge they are
The model assumes that less than some fraction of
now more focusing on homogeneous (similar in format of
the notaries act in Byzantine way [3];
contracts, accounts and consensus) blockchains based on
• Sidechains / Relays - one chain is capable of read- their own engines (Tendermint in Cosmos and Substrate1
ing and validating events on other chains. Tech- in Polkadot).
nique that allows chains to act based on events from
another ledger with the help of a smart contract. Other solutions are also suffering the same issue and a
They are a more direct approach for chain interop- lot of more ones. Blocknet [6] focuses only on solving
erability without trusted intermediate participants problems with cryptocurrency assets, concretely making
as opposed to notary schemes [3]. They usually need a fully decentralized exchange. BTC Relay [5] concen-
copies of block headers from the other chain. These trates only on one kind of interconnection, specifically
block headers usually contain Merkle Root tree hash connecting Ethereum and Bitcoin blockchains. Corda [12]
by which any transaction can be proven to be con- is not blockchain in its fundamentals. It has some at-
tained in the block from the other chain [16]. Relay tributes of blockchain technology but tries to solve all
schemes are quite powerful because they allow for problems only with notary schemes. Even if Corda was
asset portability, atomic swaps or other more com- blockchain, it would not be able to make a connection be-
plex use case without restrictions [3]. The draw- tween existing blockchain networks because Corda needs
back is that the underlying blockchains one in case to have specifically implemented systems to be able to
of one-way and both in case of two-way relays needs communicate with each other. InfiniteChain [14] intro-
to support such smart contracts; duces only sidechains to one blockchain network to make

• Hash-locking - transactions on two chains have 1

a framework for building decentralised systems;
the same activation, usually a disclosure of some https://www.parity.io/substrate/
10 Košt’ál, K.: Multi-Chain Architecture for Blockchain Networks

Table 1: Cross-chain Techniques and Their Use Cases

Hash-locking Notary Sidechains /
schemes Relays
Atomic Yes Yes Yes (only
swaps 2-way)
Asset No Yes (requires Yes
portability long-term
notary trust)
Cross-chain No Yes Yes
oracles
Asset No Yes (requires Yes
encumbrance long-term
notary trust)

it more private, faster, scalable and distribute high data nication appeared. However, despite the large number
volumes. Thus it cannot connect more existing blockchain of use cases and attempts to address them, the under-
networks. Interledger Protocol (ILP) [13] is a protocol lying problem of interblockchain communication has not
for communication between any ledgers, not only oriented been clearly defined, nor have the related challenges or
into blockchains but also to ledgers held by existing banks. the existing research [29] been studied. The lack of gen-
However, as the ILP uses Hash Timed Locks, it man- erality and dependence on the specific implementation of
dates that all sides of communication implement con- blockchains are disadvantages known to the vast major-
tracts. The Hyperledger Quilt [15], side by side with ity of existing solutions. All blockchain solutions work on
Ripple, is just a real implementation of Interledger proto- their ”sand” and are not interested in communication be-
col into blockchain technology, so it also shares the same tween a huge number of blockchains and the interconnec-
drawbacks as ILP. Lightning Network [23] is comparable tion of existing solutions. None of the options described
to InfiniteChain but it is not a sidechain. A sidechain above support plug-in connection of existing blockchains.
counts on its own blockchain. This network is, with the Only some of the solutions may comply with some gov-
help of a two-way peg, coupled to the existing blockchain. ernment regulations or may communicate with standard
On the other hand, the Lightning network consists of na- databases. In addition, some of the designs add additional
tive Bitcoin 2-of-2 multi-sig transactions. The cross-chain overheads to already slow and unpredictable blockchain
communication with Lightning Network works as long as networks, which only limits functionality. So there are
all the chains support the same hash function to use for many gaps in interoperability that need to be addressed.
the hash lock. Nodes cooperating in the Lightning Net-
work have to be always online in order to send and re- 3. Research Goals
ceive payments, and this lowers the security as cold wal- According to the references and analyzed solutions, we
lets cannot be used. Rootstock [18] is the next example certainly see room for improved interoperability between
of sidechain, but it has a significant disadvantage, that existing blockchain networks not only specialized in cryp-
it is suited only for Bitcoin blockchain. Despite that, its tocurrencies, but also databases or perhaps government
only use is to make faster transactions with the help of and state technologies.
Simplified Payment Verification2 . TAST project [1] is still
in development and not finished, but their approach with All our research is focused on the use of modern blockchain
tokens in all blockchains requires that each blockchain techniques and related technologies in combination with
has smart contracts functionality to create the tokens. our knowledge of networking and architecture. Summary
Moreover, in order to function, all wallet balances and of the main goals:
movable tokens must be in all participating blockchains.
This can radically higher storage demands. Wanchain
[19], with the help of Ethereum smart contracts, makes • Introduce a new multi-chain blockchain archi-
an interconnection between different Decentralized Apps tecture to connect two or more different blockchains
(DApps), but for now, it works only with Ethereum-based to allow interoperability between them.
blockchains. XClaim [31] is not a universal solution for
blockchain interoperability. By far, it was implemented • Enable attach existing blockchains in an almost
only between Bitcoin and Ethereum. The drawbacks are plug & play manner without the need to implement
the need to have at least one smart contracts compati- specific requirements on their part.
ble blockchain, for each pair of blockchains you need to • Introduce real-world asset backing mechanism
implement specific smart contract only for that pair, and in token-driven blockchain, e.g. a gold-protected to-
the whole trust in transactions is based on a vault, which ken as a store of value to prevent volatility.
locks the funds.

In all three theses, it is important to think about secu-

2.3 Summary rity and, where necessary, ensure that assets can be ex-
A number of competing projects presented here aimed at changed between blockchains (for example, in the case
creating a unified platform for inter-blockchain commu- of cryptocurrency volatility, redeeming an asset from the
real world with a more stable value) and ensuring that this
2 problem can be solved without a centralized third party,
https://en.bitcoinwiki.org/wiki/Simplified_
Payment_Verification e.g. an exchange. The standard approach to atomic swaps
 Information Sciences and Technologies Bulletin of the ACM Slovakia, Vol. 12, No. 2 (2020) 8-14 11

between blockchain chains is currently hashed time-locked 4.2 Color Blockchain

contracts and smart contracts, which function as two-way An architecture is proposed to solve interoperability prob-
relays. lems between current blockchain networks. The proposed
design introduces a multi-chain architecture for new, as
4. Architecture Design well as existing blockchains. We called the architecture
In the proposal to solve the described problems we have color blockchain. We do not provide any details or in-
two parts. The first is a unique double blockchain technol- structions regarding implementation, they are up to the
ogy, where one is public and contains regular transactions, developers. Examples of such details are the use of Merkle
and the second is private, which stores information about trees, specific consensus algorithms, cryptography, keys,
the real coverage of the assets from the first blockchain, and other blockchain bases that are inherited from the
such as gold. The second part of the design is in the blockchains on which applications are built. Some basic
whole network ecosystem for new blockchains and also concepts of architecture are given here.
for existing ones. We decided to think of the blockchain
as a color, so each blockchain has its own color assigned 4.2.1 Problem with "Too Large" Data
to it. That’s why we present the color blockchain. The The coexistence of color blockchains can lead to a huge
network is controlled by colors grouped into color spaces. number of generated blocks during the life of the network.
The network topology is logically structured as a network To relieve the pressure on the required performance and
of clearly colored networks, as shown in the Figure 1. storage, we need the Epoch concept based on light clients
[30] and Merklix tree3 .
4.1 Double-blockchain
For double blockchain purposes, we will use authorized 4.2.2 Participation in Colors and Color Space
blockchains, authentication will only be possible through Each block belongs to exactly one color. This color be-
the authentication block, and the token register in the longs to the color space, and there is a tree hierarchy of
authentication subsystem will remain private. The vali- color spaces.
dation block is the connection point between the public Nodes participating in a particular color usually process
and private blockchain. The authentication block can be only blocks of that color. They can (but are not obliged
an application interface (API) or a smart contract. This - depending on the rules in the color space) also process
option represents the detail of the implementation, both blocks belonging to the nearest / smallest color space to
cases should work without problems, but the API is an which they belong. Zero color blocks in color spaces are
option that can be implemented faster. In order to be able mainly used for color space management and inter-colored
to verify the token validity status in a trouble-free man- communication.
ner, i.e. validation will take place after the transaction is
signed without significant delay in the speed of the trans- 4.2.3 Color Blockchain Custom Network
action, we will need a fast scalable consensus algorithm. The color blockchain network is a complete blockchain
The ideal consensus algorithm seems to be the Practical technology that can handle thousands of individual block-
Byzantine-Fault Tolerance (PBFT) used in Hyperledger chains and a wide range of configurations. As part of the
Fabric, which can reach more than 20,000 transactions per introduction, the network will create two custom block-
second (TPS) [9]. However, to complement the results of chain implementations on the network: Color Blockchain
the mentioned article, their testing was performed on Fab- and Color Verification Double Blockchain. Both of these
ric 1.2 and only with a critical path, not with complete blockchains can be realized in separate color spaces as
nodes. According to the authors, the results will be about individual colors. This color space can have a special
25-30 % lower for complete nodes. Double-blockchain is customized set of rules that validates the properties of
an exact solution to my goal number three. tokens in one color using properties in another color.

4.2.4 Smart Smart Contracts

Smart contract design allows to use color blockchain as
a robust execution environment with pluggable environ-
ments and contract configuration. Smart Contract Exe-
cution Engine isolates the execution of smart contracts in
a quarantine environment, and therefore protects the net-
work from malicious behavior and increases security [7].

Finally, the key features of the architecture are:

• Blockchain colors - the color blockchain network is

designed from the ground up based on the idea of
partial blockchains. It can be divided into separate
sub-blockchains with their own rules and behavior.
• Enormous readiness - the networks is designed to al-
low a huge number of operations in terms of through-
put and overall capacity thanks to its concept of
distributed processing and blockchain snapshots.

3
https://www.deadalnix.me/2016/11/06/
Figure 1: Color blockchain network. using-merklix-tree-to-shard-block-validation/
12 Košt’ál, K.: Multi-Chain Architecture for Blockchain Networks

• Double-blockchain - in the center of color blockchain

is the concept of double blockchain transaction ver-
ification. This can be applied to support mech-
anisms, e.g. gold-covered assets, cryptocurrency-
covered assets, etc.

• Attaching existing blockchains - the only prerequi-

site for communicating with existing blockchains is
the API. The API can be provided directly by the
target blockchain, some API service for blockchains,
or the developer can build it himself.

These are the key concepts of architecture. If they work,

the whole architecture works like a monolith. Colors are
the basic thing about an architecture that provides in-
formation about transactions, interoperability, but also Figure 2: Experimental topology.
makes the system easier to understand. The second fea-
ture concerns scalability. Slow blockchains already exist,
but fast blockchains with a very high number of transac- test 100 times and the results shown here are the median
tions per second are, as the overview shows, very current values.
and academia and industry are trying to reach new mile-
stones. Double-blockchain is a very innovative feature The scenario was to verify the speed of transactions within
that shows other use cases where blockchain can be used. the color blockchain network. When we compare it with
This technology can be selected for any combination of the Tendermint from which this blockchain is built, we
blockchain technologies (public, private). The combina- get very similar results, almost 20,000 TPS. The results
tion of existing blockchains and distributed ledger tech- are shown in the Figure 3.
nologies, i.e. Heterogeneous interoperability is the highest
priority in this work and also a current research topic in 5.2 Summary
the field of inter-blockchain communication. This chapter described the verification of the proposed
architecture. The verification started with the setting up
5. Experimental Verification of networks and processes, which also served as a basic
We decided to implement a smaller grain of the proposed verification for defined transactions and blocks. Verifi-
solution to see the key aspects of the architecture and cation with experimental implementation followed. Both
to demonstrate the key concepts behind the design in a selected methods are independent of each other and con-
simple way. firm the functionality of parts of the proposed architec-
ture. In the experimental case, we verified the design
by implementing part of the architecture and the sample
5.1 Experimental Methodology network using several different blockchains. We divided
For experimental purposes, we implemented the topology this case into two scenarios. One where we focused on
shown in the Figure 2. The decision about colors and double blockchain technology to point out its speed and
color spaces was random without special significance. We minimal impact on extension when verifying gold sup-
have grouped related blockchains with respect to their port for each transaction. The test showed minimal dif-
consensus algorithm used, but note that this is only cur- ferences in times with and without validation, the only
rent choice for experiments. case where the difference was more seen, i.e. 831 ms,
was with 50,000 transactions, but so many transactions
In this experimental topology, we use APIs provided from usually do not occur in one blockchain network at once.
blockchains. However, if the blockchain does not provide In the second scenario, we verified the partially imple-
its own API, we must use some service that the API will mented color blockchain architecture using three existing
create or make available. For this service we use Cryp- blockchain networks, a double-blockchain, and six color
toAPIs 4 , which is an infrastructure layer that provides blockchain-derived blockchains. In addition, there were
APIs, WebHooks and Web Sockets for some of the ex- APIs to connect to existing blockchains. The results in
isting blockchain protocols, e.g. Litecoin, Bitcoin, Dash, this scenario show that our network is high-speed and its
Dogecoin and more. slowdown is only affected by the length of the verification
time in the source and destination blockchain. Subse-
The blockchains selected for testing were the public test- quently, we verified the scalability on blockchains derived
ing networks Bitcoin, Ethereum and Eos. These three from color blockchain, which are not significantly slowed
existing blockchain networks are sufficient to demonstrate down by block time, and here it turned out that the en-
functionality and communication with existing block- tire color blockchain network managed up to 18,000 TPS.
chains. Color blockchain uses the Tendermint core [2]. During the network latency and throughput test scenar-
ios, we monitored the impact of our implementation and
All experiments were performed on Amazon EC2 instances use of consensus algorithms on network properties. Mea-
of type m5n.4xlarge, where we ran the tests with 4, 8, surements were performed in a real geographically dis-
16 and 32 instances. Each was a validator for the core tributed environment to indicate network operation if de-
network and some were validators and vaults for exist- ployed. The real environment was set so that its prop-
ing blockchains and double-blockchain. We repeated each erties do not affect the quality. In all experiments, we
verified that the color blockchain network has a minimal
4
https://cryptoapis.io/products/blockchain-apis/ impact on the quality of services provided.
 Information Sciences and Technologies Bulletin of the ACM Slovakia, Vol. 12, No. 2 (2020) 8-14 13

Figure 3: Delay vs. throughput in a testing environment.

6. Conclusion [2] E. Buchman, J. Kwon, and Z. Milosevic. The latest gossip on BFT
In this work, we presented a solution to problems with the consensus. arXiv, 1807(4938):14, jul 2018.
interoperability of various blockchain networks. Our ar- [3] V. Buterin. Chain Interoperability. R3 Research Paper, 2016.
chitecture supports atomic swaps between existing block- [4] Z.-d. CHEN, Z. YU, Z.-b. DUAN, and K. HU. Inter-Blockchain
chain networks using vaults and collateralization. There Communication. In Proc. of the 2nd Int. Conf. on Computer
are also notary schemes. The current literature has shown Science and Technology (CST 2017), pages 448–454, Guilin,
China, 2017. DEStech Publications.
that we cannot implement blockchain interoperability
[5] L. Coleman. BTC Relay Bridges BTC With Ethereum, Allowing
without a trusted third party [29]. However, in our so-
BTC Verification For Smart Contracts, 2016.
lution, validators in a trusted third party come from par-
[6] A. Culwick and D. Metcalf. The Blocknet Design Specification.
ticipating blockchains, so even though our transmission Technical report, 2018.
chain is called ”centralized,” if we use nodes from existing [7] G. G. Dagher, C. L. Adhikari, and T. Enderson. Towards Secure
networks where they are trusted, they will also be trusted Interoperability between Heterogeneous Blockchains using Smart
in the color blockchain. Contracts. In Future Technologies Conf. (FTC) 2017, pages
73–81, Vancouver, Canada, 2017. Science and Information
The results of verifications confirmed the functionality of Conferences.
the proposed architecture and the fulfillment of the de- [8] I. Eyal. Blockchain Technology: Transforming Libertarian
fined goals. We introduced a new multi-chain blockchain Cryptocurrency Dreams to Finance and Banking Realities.
architecture for interoperability that has proven itself, so Computer, 50(9):38–49, 2017.
the first goal is complete. The second goal is also achieved [9] C. Gorenflo, S. Lee, L. Golab, and S. Keshav. FastFabric: Scaling
thanks to the API for existing blockchains and the intro- Hyperledger Fabric to 20,000 Transactions per Second. In 2019
duction of colors and color spaces. We tested the connec- IEEE Int. Conf. on Blockchain and Cryptocurrency (ICBC), pages
tion of three different blockchains to our network; linking 455–463. IEEE, 2019.
a new blockchain is a process that takes a few seconds, [10] Y. Guo and C. Liang. Blockchain application and outlook in the
banking industry. Financial Innovation, 2(24):1–12, dec 2016.
because the only requirement is to map the functions of
[11] T. Hardjono, A. Lipton, and A. Pentland. Towards a Design
the target blockchain to color blockchain. The third goal
Philosophy for Interoperable Blockchain Systems. arXiv,
is achieved using double blockchain technology. In addi- 1805(5934):27, may 2018.
tion, atomic swaps are achieved in the architecture using [12] M. Hearn. Corda: A distributed ledger. Technical report, 2016.
smart contracts (i.e. relay schemes) and automatic locks [13] A. Hope-Bailie and S. Thomas. Interledger: Creating a Standard
between accounts via a transmission chain (i.e. notary for Payments. In Proc. of the 25th Int. Conf. Companion on World
scheme), but the system is also ready for hash locks, this Wide Web - WWW ’16 Companion, pages 281–282, New York,
will only be an implementation detail. New York, USA, 2016. ACM Press.
[14] G.-H. Hwang, P.-H. Chen, C.-H. Lu, C. Chiu, H.-C. Lin, and A.-J.
Acknowledgements. This research was supported by Jheng. InfiniteChain: A Multi-chain Architecture with Distributed
the Ministry of Education, Science, Research and Sport Auditing of Sidechains for Public Blockchains. In S. Chen et al.,
of the Slovak Republic, Incentives for Research and Devel- editors, Lecture Notes in Computer Science (including subseries
opment, Grant No.: 2018/14427:1-26C0. This publication Lecture Notes in Artificial Intelligence and Lecture Notes in
was created thanks to support under the Operational Pro- Bioinformatics), volume 10974 LNCS, pages 47–60. Springer
gram Integrated infrastructure for the project: Research International Publishing, Cham, jun 2018.
in the field of blockchain technology with connection to [15] Hyperledger Inc. Hyperledger Gets Cozy With Quilt, 2017.
online payment services, ITMS 313022U641, co-financed [16] T. Koens and E. Poll. Assessing interoperability solutions for
distributed ledgers. Pervasive and Mobile Computing,
by the European Regional Development Fund. It was
59(101079):1–10, oct 2019.
also partially supported by the grants APVV-15-0731,
[17] J. Kwon and E. Buchman. Cosmos: A Network of Distributed
KEGA 011STU-4/2017, and VEGA 1/0836/16. The au- Ledgers. Technical report, 2018.
thor would like to thank for financial contribution from
[18] S. Lerner. Drivechains, Sidechains and Hybrid 2-way Peg
the STU Grant scheme for Support of Young Researchers. Designs. Technical report, RSK Labs Ltd., 2016.
[19] J. Lu, B. Yang, Z. Liang, Y. Zhang, S. Demmon, E. Swartz, and
References L. Lu. Wanchain: Building Super Financial Markets for the New
[1] M. Borkowski, D. Mcdonald, C. Ritzer, and S. Schulte. Towards Digital Economy. Technical report, 2017.
Atomic Cross-Chain Token Transfers: State of the Art and Open [20] M. Mettler. Blockchain technology in healthcare: The revolution
Questions within TAST. Technical report, TU Wien, Vienna, starts here. In 2016 IEEE 18th Int. Conf. on e-Health Networking,
Austria, 2018. Applications and Services (Healthcom), pages 1–3. IEEE, 2016.
14 Košt’ál, K.: Multi-Chain Architecture for Blockchain Networks

[21] Q. K. Nguyen. Blockchain - A Financial Technology for Future [30] A. Zamyatin, Z. Avarikioti, D. Perez, and W. J. Knottenbelt.
Sustainable Development. In 2016 3rd Int. Conf. on Green TxChain: Efficient Cryptocurrency Light Clients via Contingent
Technology and Sustainable Development (GTSD), pages 51–54. Transaction Aggregation. Technical report, Cryptology ePrint
IEEE, nov 2016. Archive, 2020.
[22] L. Pawczuk, J. M. Nielsen, P. K. H. Sin, and N. Hewett. Inclusive [31] A. Zamyatin, D. Harz, J. Lind, P. Panayiotou, A. Gervais, and
Deployment of Blockchain for Supply Chains: Part 6-A W. Knottenbelt. XCLAIM: Trustless, Interoperable,
Framework for Blockchain Interoperability In Collaboration with Cryptocurrency-Backed Assets. In 2019 IEEE Symposium on
Deloitte. Technical report, Deloitte, 2020. Security and Privacy (SP), pages 193–210. IEEE, may 2019.
[23] J. Poon and T. Dryja. The Bitcoin Lightning Network: Scalable
Off-Chain Instant Payments. Technical report, 2016.
[24] M. Swan. Blockchain : blueprint for a new economy. O’Reilly Selected Papers by the Author
Media, 1 edition, 2015. K. Košt’ál, P. Helebrandt, M. Belluš, M. Ries, and I. Kotuliak.
[25] S. Underwood. Blockchain beyond bitcoin. Communications of Management and Monitoring of IoT Devices Using Blockchain.
the ACM, 59(11):15–17, oct 2016. Sensors, vol. 19, no. 4, p. 856, Feb. 2019.
[26] H. Wang, Y. Cen, and X. Li. Blockchain Router. In Proc. of the K. Košt’ál, T. Krupa, M. Gembec, I. Veres, M. Ries, and I. Kotuliak.
6th Int. Conf. on Informatics, Environment, Energy and On Transition between PoW and PoS. In Proceedings of 2018
Applications - IEEA ’17, pages 94–97, New York, New York, International Symposium ELMAR: 60th International
USA, 2017. ACM Press. symposium., pages 207–210, Zadar, Croatia, 2018. IEEE Press.
[27] G. Wood. POLKADOT: VISION FOR A HETEROGENEOUS V. Valaštín, K. Košt’ál, R. Bencel, and I. Kotuliak. Blockchain Based
MULTI-CHAIN FRAMEWORK. Technical report, 2017. Car-Sharing Platform. In Proc. of ELMAR-2019: 61st Int.
[28] J. Yli-Huumo, D. Ko, S. Choi, S. Park, and K. Smolander. Where symposium., pages 5–8, Zadar, Croatia, 2019. IEEE Press.
Is Current Research on Blockchain Technology?-A Systematic
K. Košt’ál, R. Bencel, M. Ries, and I. Kotuliak. Blockchain E-Voting
Review. PLOS ONE, 11(10):163–189, oct 2016.
Done Right: Privacy and Transparency with Public Blockchain.
[29] A. Zamyatin, M. Al-Bassam, D. Zindros, E. Kokoris-Kogias, In 2019 IEEE 10th International Conference on Software
P. Moreno-Sanchez, A. Kiayias, and W. J. Knottenbelt. SoK: Engineering and Service Science (ICSESS), pages 592–592.
Communication Across Distributed Ledgers. Technical report, Beijing, China, 2019. IEEE Press
Cryptology ePrint Archive, 2019.
Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.