Conference PaperPDF Available

Abstract and Figures

Distributed Ledger Technology (DLT) enables a new way of inter-organizational collaboration via a shared and distributed infrastructure. There are plenty of DLT designs (e.g., Ethereum, IOTA), which differ in their capabilities to meet use case requirements. A structured comparison of DLT designs is required to support the selection of an appropriate DLT design. However, existing criteria and processes are abstract or not suitable for an in-depth comparison of DLT designs. We select and operationalize DLT characteristics relevant for a comprehensive comparison of DLT designs. Furthermore, we propose a comparison process , which enables the structured comparison of a set of DLT designs according to application requirements. The proposed process is validated with a use case analysis of three use cases. We contribute to research and practice by introducing ways to operation-alize DLT characteristics and generate a process to compare different DLT designs according to their suit-ability in a use case.
Content may be subject to copyright.
Do Not Be Fooled: Toward a Holistic Comparison of
Distributed Ledger Technology Designs
Florian Gräbe
Karlsruhe Institute of
Technology
florian.graebe
@web.de
Niclas Kannengießer
Karlsruhe Institute of
Technology
niclas.kannengiesser
@kit.edu
Sebastian Lins
Karlsruhe Institute of
Technology
lins@kit.edu
Ali Sunyaev
Karlsruhe Institute of
Technology
sunyaev@kit.edu
Abstract
Distributed Ledger Technology (DLT) enables a
new way of inter-organizational collaboration via a
shared and distributed infrastructure. There are plenty
of DLT designs (e.g., Ethereum, IOTA), which differ in
their capabilities to meet use case requirements. A
structured comparison of DLT designs is required to
support the selection of an appropriate DLT design.
However, existing criteria and processes are abstract
or not suitable for an in-depth comparison of DLT de-
signs. We select and operationalize DLT characteris-
tics relevant for a comprehensive comparison of DLT
designs. Furthermore, we propose a comparison pro-
cess, which enables the structured comparison of a set
of DLT designs according to application require-
ments. The proposed process is validated with a use
case analysis of three use cases. We contribute to re-
search and practice by introducing ways to operation-
alize DLT characteristics and generate a process to
compare different DLT designs according to their suit-
ability in a use case.
1. Introduction
Given the high potential of distributed ledger tech-
nology (DTL), numerous DLT designs have been de-
veloped during the past decade (e.g., Ethereum, IOTA,
and Tezos). Such DLT designs specialize in fulfilling
requirements of a particular set of applications on DLT
in domains such as the Internet of Things (IoT), fi-
nance, and supply chain management. However, this
specialization resulted in trade-offs between DLT
characteristics that restrict the suitability of a DLT de-
sign to a particular set of applications [1, 2]. For ex-
ample, a DLT design cannot provide high availability
and a high degree of consistency simultaneously be-
cause a large number of nodes of the ledger, which is
needed for high availability, requires more time and
effort to be synchronized, thereby challenging con-
sistency [1, 2]. Due to the prevalent trade-offs between
DLT characteristics, the improvement of one DLT
characteristic deters another, inhibiting a one-size-fits-
all DLT design suiting all requirements of individual
applications. Trade-offs thus require developers to
choose the best fitting DLT design for their application
[2]. Making careful and well-founded decisions in fa-
vor for an (appropriate) DLT design to develop viable
applications on DLT becomes even more crucial be-
cause technical differences between DLT designs
(e.g., different data structures and consensus mecha-
nisms) impede the ex-post migration of data stored on
one distributed ledger to another [3].
Developers, therefore, need to conduct a compre-
hensive comparison between prospective DLT designs
to assess DLT designs’ suitability for a particular ap-
plication on DLT before starting the actual implemen-
tation. Such comparisons require the operationaliza-
tion of DLT characteristics that is referred to as a pro-
cess of defining the measurement of DLT characteris-
tics to make them understandable, measurable, and
comparable. However, it remains challenging for de-
velopers to compare DLT designs and operationalize
DLT characteristics because DLT synthesizes multiple
techniques of computer science (e.g., cryptography
and distributed databases), which come with individ-
ual operationalizations for characteristics. In addition,
DLT exhibits unique characteristics, such as stale
block rate and smart contract support [1], requiring
new operationalizations. The operationalization of
DLT characteristics must first be clarified to enable
developers conducting comparisons of DLT designs.
Research on DLT in computer science has already
taken different perspectives on the analysis, operation-
alization, and benchmarking of DLT designs, for ex-
ample, in regard to formal verification of consensus
mechanisms (e.g., [4]) and the analysis of DLT de-
signs in different configurations (e.g., [5, 6]). How-
ever, prior research predominantly considers DLT
characteristics related to performance (e.g., [7, 8]) and,
thus, neglects further important characteristics (i.e.,
flexibility, anonymity) and does not allow a holistic
comparison of DLT designs. Research on DLT in in-
formation systems deals with the specification of pro-
cesses to support developers in their selection of an ap-
propriate DLT design for an application (e.g., [9, 10]),
among others. However, the presented processes are
often too abstract to compare DLT designs. To con-
duct a comprehensive comparison of DLT designs it
requires the synthesis of the prevalent research streams
on the operationalization of DLT characteristics and
benchmarking DLT designs, and decision support in
the selection of an appropriate design. We ask the fol-
lowing research question (RQ):
RQ: How can DLT designs be compared according
to application requirements prior to implementation?
To answer the RQ we follow two objectives: first
the identification and operationalization of relevant
DLT characteristics and, second, the development of a
process for the comparison of DLT designs. For the
first, we consolidated a set of DLT characteristics that
we deem relevant for applications on DLT and synthe-
sized existing research on benchmarks and operation-
alizations of DLT, distributed databases, and infor-
mation systems. Second, we generated and validated a
process for the comparison of DLT designs with three
prominent use cases.
With our study, we contribute to practice as we en-
able a comprehensive comparison of DLT designs
and, thus, provide decision support for the selection of
a suitable DLT design for viable applications on DLT.
We contribute to research because we synthesize ex-
isting approaches and generate new means for the op-
erationalization of DLT characteristics. Thus, our re-
sults can serve as a foundation for research on the se-
lection of suitable DLT designs for applications.
2. Related research
2.1. Distributed ledger technology
DLT serves as a shared, digital infrastructure for
applications (e.g., financial transactions) because DLT
enables the operation of an append-only database (re-
ferred to as ledger), which is distributed across multi-
ple storage devices (so-called nodes) in an untrustwor-
thy environment [11]. Each node maintains a local rep-
lication of the data stored on the ledger. An untrust-
worthy environment is characterized by the arbitrary
occurrence of Byzantine failures [12] such as crashed
or (temporarily) unreachable nodes, network delays,
and malicious behavior of nodes (i.e., issuing wrong
information). In DLT, data is appended to the ledger
through transactions and is stored in a chronologically-
ordered sequence. Each transaction contains meta-data
(e.g., receiver address, timestamp) and a digital repre-
sentation of certain, tokenized assets or program code
of a smart contract. A tokenized asset is a digital rep-
resentation of an asset (e.g., coins) in a structured data
format (token), which can be transferred using trans-
actions. When a node receives a new transaction, the
transaction is validated. Valid transactions are for-
warded to all adjacent nodes, which also validate and
forward the transaction subsequently.
Because all nodes of a distributed ledger maintain
a local replication of the ledger, all nodes must be syn-
chronized and agree on a common state of the distrib-
uted ledger to reach consistency. For this purpose, a
consensus mechanism is employed [13]. A consensus
mechanism is used to manage the negotiation between
nodes, which eventually agree on a common replica-
tion of the ledger. Once appended, data can hardly be
altered or removed anymore.
2.2. Comparison of DLT designs
The comparison of DLT designs requires taking
both a technical perspective considering DLT charac-
teristics, such as fault tolerance or throughput, and an
economical perspective considering costs and time for
software development. Therefore, we first present re-
lated research in benchmarking DLT characteristics.
Second, we present related research on supporting de-
velopers to select a suitable DLT design in order to
avoid, for example, high switching costs. Benchmark-
ing forms a measurement of a (non-productive test)
system at a specific point in time [14] and allows for
the comparison of systems in artificially created sce-
narios. To date, DLT benchmarking predominantly fo-
cuses on DLT characteristics related to performance
and security (e.g., throughput, network partitioning)
[5, 15], specific use cases [16], or private DLT designs
[17]. Still, such approaches do not allow for holistic
benchmarking of DLT designs because they do not
consider DLT characteristics such as availability or
cost. Since DLT incorporates multiple domains of
computer science (i.e., databases, cryptography, infor-
mation systems), DLT benchmarking must also con-
sider extant research in these disciplines.
Database benchmarks include characteristics such
as transaction speed and consistency. Meanwhile, in-
dustry standards have been established for database
benchmarks with a focus on transaction execution,
performance, and scalability (e.g., TPC-H). While
these operationalizations might be used for certain
DLT characteristics, database benchmarks do not
cover several unique DLT characteristics (e.g., confir-
mation latency, fault tolerance).
Extant research in cryptography already provides
operationalizations for characteristics related to hash-
ing or public-key encryption, which are substantial for
DLT. From the performance perspective, for example,
time complexity of such encryption algorithms is an
important criterion to assess. From the security per-
spective, the collision resistance of hashing algorithms
is a common criterion to validate a hashing algorithm
against pre-image attacks [18]. Although the opera-
tionalization of cryptographic procedures is useful to
measure in DLT, these operationalizations serve only
one (research) domain DLT draws from.
Related research on decision support to find an ap-
propriate DLT design is still in its infancy. There are
already approaches that support the decision of using
DLT or not (e.g., [10]) and several classifications of
DLT design have been proposed to make DLT and its
characteristics better comprehensible for developers
(e.g., [19]). However, such processes do not consider
the operationalization of DLT designs and, thus, are
only applicable to DLT on a limited scale.
3. Method
We applied a four-step research approach. First,
we reviewed the extant literature on DLT to generate
a preliminary set of characteristics, which are relevant
for DLT design comparison (e.g., [1, 22]). Second, we
searched operationalizations of the respective DLT
characteristics in extant literature. Third, we evaluated
the suitability of DLT characteristics and correspond-
ing operationalizations for the comparison process fol-
lowing well-known IT benchmarking requirements
and quality criteria for metrics (e.g., [14, 20, 21, 23];
cf. Table 1 and Table 2). Finally, we generated a DLT
comparison process and evaluated its usefulness using
three use cases and five DLT designs (i.e., Ethereum,
Hyperledger Indy, IOTA, Tezos and Hashgraph).
3.1. Selecting suitable characteristics for the
comparison of DLT designs
To identify candidate DLT characteristics for our
comparison process, we built on our prior research on
identifying and clustering DLT characteristics [1, 24,
25], leading to a set of 37 DLT characteristics. More-
over, we reviewed further research articles and white-
papers on DLT characteristics (e.g., [1, 19, 26]) and on
DLT benchmarking (e.g., [22]), which led to a final set
of 50 DLT characteristics as candidates for the com-
parison process. We classified these DLT characteris-
tics into qualitative and quantitative sub-groups to
conduct a more structured search for operationaliza-
tion approaches. As a next step, we performed a fo-
cused literature search for each DLT characteristic to
identify potential operationalizations in DLT research
and related research streams (i.e., databases, cryptog-
raphy, information systems). In particular, we
searched in scientific databases, including IEEE
Xplore, EBSCOHost, ACM Digital Library, and
Proquest. For each DLT characteristic, we created a
unique search string comprising the name of the DLT
characteristic (i.e., scalability) and synonyms for op-
erationalization (i.e., benchmarking, metrics, meas-
urement). We reviewed the resulting research articles
and noted proposed operationalizations for each char-
acteristic. The identified operationalizations can be
classified into three different categories: First, opera-
tionalizations found in the (non-) scientific literature
on DLT that could directly be inherited into the com-
parison process (e.g., operationalizations for con-
sistency or confirmation latency). Second, operation-
alizations found in related domains that needed to be
adapted to fit the DLT context, such as operationaliza-
tions for developer support, scalability, or availability.
Lastly, we were left with four DLT characteristics
(i.e., traceability, transaction content transparency),
where no applicable operationalizations were found.
For these DLT characteristics, we developed new met-
rics complying with the criteria from Table 2.
To assess whether identified DLT characteristics
are suitable for DLT design comparison, we evaluated
whether they fulfill the requirements for IT bench-
marks (cf. Table 1) and whether corresponding opera-
tionalizations comply with requirements for quality
metrics (cf. Table 2). For more information see “Ex-
clusion Methodology” in the supplementary online
material (https://bit.ly/2klPK9v). First, we investi-
gated if identified operationalizations are consistent.
Table 1. Summary of requirements for IT benchmarking [14, 20, 21]
Requirement
Description
Economic Efficiency
It must be economically affordable to run the benchmark.
Fairness
Fairness means that a benchmark should treat every system under test fairly and equally and should not make assumptions
on the system’s features.
Portability
It should be easy to implement the benchmark on many different systems and architectures.
Relevance
The benchmark should focus on typical operations within that problem domain.
Reproducibility
Reproducibility implies that running the same benchmark multiple times will yield similar results, meaning that a certain
degree of determinism is required.
Simplicity
The benchmark must be understandable, otherwise, it will lack credibility. A key aspect is the presence of meaningful and
expressive metrics.
Scalability
The benchmark should apply to small and large computer systems. It should be possible to scale the benchmark up to larger
systems and to parallel computer systems as computer performance and architecture evolve.
We excluded 17 DLT characteristics (i.e., governance,
non-repudiation, interoperability) because a consistent
output of identified operationalizations could not be
guaranteed as, for instance, the operationalizations re-
quire a subjective view on qualitative DLT character-
istics. For example, we excluded the DLT characteris-
tic compliance because a metric that operationalizes
compliance would have a strongly varying output de-
pending on the particular regulations, thereby threat-
ening consistency. As a next step, we excluded five
DLT characteristics (e.g., incentive mechanism) that
did not follow the high-resolution criteria of quality
metrics (cf. Table 2). The remaining DLT characteris-
tics fulfill the requirements for quality metrics, includ-
ing correlation, discriminative power, and tracking. As
the last step, we examined if the final set of DLT char-
acteristics fulfills all requirements for IT benchmark-
ing (cf. Table 1). We removed eight DLT characteris-
tics that did not comply with the relevance criteria,
leading to a final set of 20 relevant DLT characteristics
for the comparison process. For example, block crea-
tion interval was excluded because it is expressed in
throughput and stale block rate [27], which can be di-
rectly mapped to application requirements.
3.2. Comparison process development
We grounded the development of a comparison
process on previous research of benchmarking pro-
cesses in the IT field (e.g., [28, 29]). We also drew
from extant research on benchmark process develop-
ment (e.g., [29]) and adapted it to the field of DLT un-
der consideration of the requirements depicted in Ta-
ble 1. We used [29] as a basic approach for a bench-
marking process and adapted it in two discussion
rounds with researchers to fit the DLT context while
complying with the requirements from Table 1. Dur-
ing this adaptation, the important core steps of the
comparison process were identified, adapted and en-
hanced to our use case. The individual steps were then
renamed to keep consistent terminology. We showed
the usefulness of the comparison process by evaluating
the process in three prominent use cases in the field of
DLT: cryptocurrency, IoT, and identity management.
We selected these use cases because they form a high
percentage of identified DLT use cases in research
[30] and currently worked on by large companies [31].
For each use case, we applied the comparison process
and included five strongly different DLT designs,
which have been developed for different purposes:
Ethereum as a general-purpose blockchain, Hashgraph
as a public multi-purpose DLT design, IOTA with a
focus on IoT, Hyperledger Indy for identity manage-
ment, and Tezos for strong governance to ensure a
wide variety of different ledgers and to show the use-
fulness of the process for different DLT designs. We
chose these DLT designs because they strongly differ
in the used data structures (e.g., Hashgraph and IOTA
follow the concept of transaction-based, directed acy-
clic graph instead of a linked chain of blocks like in
Ethereum, Hyperledger, and Tezos), the applied con-
sensus mechanisms (e.g., IOTA uses the Tangle and
Ethereum relies on Proof-of-Work), and in their per-
missioning (Ethereum as a permissionless DLT de-
signs and Hyperledger Indy as permissioned DLT de-
sign). Thus, we show that the developed operationali-
zations are applicable to a variety of DLT designs and
reflect their individual strengths.
4. Results
4.1. DLT characteristics for the comparison
4.1.1. Flexibility. Flexibility incorporates the possibil-
ities offered by a DLT design for maintenance and fur-
ther development [1]. The purpose of tokens (F1) can
be classified into achieving three different objectives:
payment tokens (e.g., Bitcoin), utility tokens (e.g.,
Filecoin), and security tokens (e.g., tZERO). Smart
contract support (F2) is a qualitative characteristic that
can be mapped to a nominal scale: either a DLT design
offers Turing-complete, Turing-incomplete, or no
smart contracts.
4.1.2. Institutionalization. Institutionalization de-
scribes the embedding of concepts and artifacts (here
DLT) in social structures. Developer support (I1) is
mainly driven by the size of the developer community
currently dealing with a particular DLT design. There-
fore, we operationalize developer support as the num-
ber of active developers. A developer is considered ac-
tive if she has at least one commit to a DLT design
core (e.g., Ethereum Protocol) or a project related to a
Table 2. Summary of requirements for quality metrics [14, 23]
Requirement
Description
Consistency
The output of the metric should be consistent. A monotonic function is often required.
Correlation
There should be a (not necessarily linear) correlation between the dimension under observation and the metric output.
Discriminative
Power
The metric should accurately display differences of parameter levels and especially differentiate between high and low pa-
rameter levels.
High resolution
The metric should have a large number of possible output values and should avoid unnecessary aggregation.
Tracking
The metric should build on the current state of the system.
DLT design (e.g., an application using the DLT de-
sign) on GitHub during the last three months. Liability
(I2) is classified as a qualitative characteristic that we
map to a nominal scale with binary values existence of
or non-existence of an enterprise organization for the
purpose of operationalization.
4.1.3. Anonymity. Anonymity describes the degree to
which individuals are not identifiable within a set of
users [1]. Node verification (LA1) is predominantly
concerned with granting or revoking permissions for
nodes. Thus, we mapped node verification to the DLT
design types public or private to consider read permis-
sions. For write permissions, we distinguish between
permissioned and permissionless DLT designs.
For traceability (LA2), we chose three distinct val-
ues: publicly-viewable transfers (e.g., in Bitcoin); ob-
fuscated transfers (e.g., using mixing [32]); not trace-
able transfers (e.g., in ZCash). For transaction content
transparency (LA3) we chose a binary value that dis-
tinguishes between data stored in plain text or en-
crypted.
4.1.4. Performance. Performance includes DLT char-
acteristics regarding the accomplishment of a given
task measured against standards of accuracy, com-
pleteness, costs, and speed [1] such as confirmation la-
tency, throughput, and scalability. Confirmation la-
tency (P1) refers to the period required to append a
minimum number of blocks to the distributed ledger
that must at least succeed a certain block b to assure
that b cannot be altered anymore (e.g., [33]). We oper-
ationalize confirmation latency (P1) as duration until a
transaction is seen as finalized.W
For scalability (P2), we focus on horizontal scala-
bility (cf. [17]), which is operationalized by changes
to throughput (P3) and mean transaction latency
(MTL) when the number of validating nodes (de-) in-
creases. For operationalization, we use p-scalability
[34] (cf. formula 1), which compares two differently
scaled system levels k1 and k2 according to the power
metric (cf. formula 2) and scaling cost [34]. We relay
this concept to DLT by excluding the cost factor and
replacing the mean delay with the mean transaction la-
tency (MTL) (cf. [17]).
Table 3. DLT characteristics and their operationalization following [1, 8, 22]
*
DLT Chars.**
Definition
Operationalization
Flexibility
Purpose of the To-
kens (F1)
The purpose and flexibility in usage of the provided tokens
of a DLT design.



Smart Contract Sup-
port (F2)
The level of how well smart contracts are supported by a
DLT design.



Institutional.
Developer Support
(I1)
The existence of documentation, interaction platforms such
as forums, or direct contact to the team developing the dis-
tributed ledger for questions and issues related to the devel-
opment of applications integrating DLT as well as program-
ming tools and interfaces.

Liability (I2)
The existence of an organization that is responsible for the
maintenance of a DLT design.


Anonymity
Node verification
(LA1)
The extent to which new nodes can join the distributed
ledger without being verified.




Traceability (LA2)
The level to which the transfer of an asset can be traced
chronologically on a distributed ledger.



Transaction Content
Transparency (LA3)
The ability to publicly view an account’s holdings and trans-
actions’ contents on a distributed ledger.
 

Performance
Confirmation La-
tency (P1)
The average time until enough blocks (or transactions) are
added to the distributed ledger so that the likelihood of tam-
pering of a previously added block or transaction is below a
certain threshold.


Scalability (P2)
The capability of a DLT design to handle an increasing
amount of workload or its potential to be enlarged to accom-
modate that growth.



,
Throughput (P3)
The number of transactions validated and appended to the
ledger in a given time interval.


*
* DLT Property ** DLT Characteristics
(1)




(2)


In (1), throughput behaves proportionally to scala-
bility, while transaction latency behaves anti-propor-
tionally to scalability. For example, if a DLT scales up
well, one would expect throughput to increase and
transaction latency to decreases. This behavior is rep-
resented in our metric by placing throughput in the nu-
merator of the metric and the transaction latency in the
denominator. To form a score for scalability, we com-
pare two different horizontal scaling levels k1 and k2
with different numbers of participating nodes as our
metric. We assume MTL to never converging to zero.
Usually, a scalability coefficient equal to or larger than
1 reflects good scalability, while a scalability coeffi-
cient close to zero indicates bad scalability. The quo-
tient of the data from the actual period is compared to
the performance data quotient from the previous pe-
riod to form the scalability score which is then com-
puted into a final scalability score (P2).
4.1.5. Security. Security refers to the preservation of
confidentiality, integrity, and availability of infor-
mation. Availability (S1) is operationalized as proba-
bility and considers the Meant Time To Failure
(MTTF) and Mean Time To Repair (MTTR) [35].
MTTF is the period a distributed ledger is expected to
correctly operate. MTTR is the required period to re-
cover the distributed ledger from a failure. This sum is
equal to the Mean Time Before Failure (MTBF) [35].
In the field of DLT, we refer MTTR and MTTF to the
full distributed ledger instead of particular nodes.
Consistency (S2) refers to storing the identical rep-
lications of the ledger on each node at the same time
[36]. In Hyperledger Caliper, transaction latency is a
metric to measure consistency as the period for a trans-
actions effect to be available on all nodes [8]. We
adopt this interpretation and measure the period be-
tween transaction issuance and transaction confirma-
tion. We include it as a criterion in form of an average
of N measurements at different times and states of the
distributed ledger. Fault tolerance (S3) is the ability of
a distributed ledger to correctly operate in the presence
of failures [35]. We operationalize fault tolerance as
the changes in throughput and transaction latency (TL)
during node failure (cf. [17]). Node failure is the stop-
ping of a node (crash failure), network delay, or ran-
dom responses due to corrupted messages [17].
Level of decentralization (LoD) (S4) expresses the
ratio of the number of independent validating nodes
(VNs) and the number of validating node operators
(VNOs). While VNs represent validating nodes,
VNOs represent an individual or organization, who
maintains the VNs (e.g., a mining pool). To include
permissioned DLT designs in our operationalization,
Table 3 cont. DLT characteristics and their operationalization following [1, 8, 22]
*
DLT Chars.**
Definition
Operationalization
Security
Availability (S1)
The probability that a system is operating correctly at an arbi-
trary point in time.



Consistency (S2)
The state that all nodes store the same data in their ledger at the
same time.


  
Fault Tolerance (S3)
The ability of a distributed ledger to correctly operate in the
presence of (hardware or software) failures.


Level of Decentrali-
zation (S4)
The number of independent node controllers participating in
transaction validation and consensus finding.

 
Network Size (S5)
The number of validating nodes of a distributed ledger that
keep a full replication of the ledger.

Reliability (S6)
The period of time during which a system is correctly function-
ing.


Stale Block Rate
(S7)
The number of blocks in a defined period of time that has been
mined but not added to the distributed ledger. A stale block
forms a fork until it is resolved by the DLT design’s fork reso-
lution rule.


Strength of Encryp-
tion (S8)
The level of security of the applied cryptographic approach.

Usability
Censorship Re-
sistance (U1)
The equal right of any user of the distributed ledger to submit
transactions that are not altered or dropped by a third party.



Cost (U2)
Costs related to the implementation and usage of a DLT design,
including software development and operational costs.

 



* DLT Property ** DLT Characteristics VN: Validating Node VNO: Validating Node Operator
we multiply the ratio of VNs to VNOs with the number
of VNs to correctly scale it to the network size and to
differentiate between permissioned DLT designs with
a lower LoD score and permissionless networks with
a higher LoD score due to a higher number of VNOs.
Network size (S5) describes the number of nodes in
a distributed ledger that stores a full replication of the
ledger, considering only full network nodes that keep
a full replication of the ledger [1]. Thus, we operation-
alize network size as the number of full nodes in a dis-
tributed ledger.
Reliability (S6) refers to the period during which a
distributed ledger is correctly functioning. In DLT, the
component that may fail refers to the complete distrib-
uted ledger [35]. The operationalization yields a prob-
ability that the system produces a correct output up to
a time t [35]. For our benchmarking process, we intro-
duce the estimated project duration t for all DLT de-
signs and require a certain probability. Due to the little
occurrences of system failures of DLT designs, this
metric is assumed to be close to one, even for larger t.
Stale blocks impact security and performance be-
cause they lead to inconsistency between nodes
through forks. The stale block rate (S7) is operation-
alized as a percentage of the mined but not included
blocks measured over the last 1000 blocks [7]. Non-
forking consensus mechanisms (e.g., Practical Byzan-
tine Fault Tolerance) get the score of zero.
The strength of encryption refers to the level of se-
curity concerning the application of authentication-re-
lated cryptographic primitives (e.g., hashing algo-
rithm). For the benchmarking process, we use the col-
lision resistance of the applied hashing algorithm in
the DLT design, which refers to the ease in guessing a
pre-image for a hash value. A Birthday attack is a
probabilistic approach of guessing pre-images exploit-
ing the fixed degree of permutations [37]. A Birthday
attack evaluates a hash function with n input bits and
m output bits for randomly-selected inputs until two
matching outputs are found. The number of pairs (p)
in-between inputs which may yield to a collision
grows quadratically with the number of trials l in a
Birthday attack (cf. formula 3). As every pair of inputs
is a chance for a matching output, finding a collision
becomes more and more likely. Using a Birthday at-
tack, it is possible to find a collision of two different
inputs with Formula 4 or better time attack (tattack) [37]
leading to Formula 5 as the security in (output-) bits
(sbit) of h.
(3)

(4)

(5)

4.1.6. Usability. Usability refers to the extent to which
a DLT design can be used by specified users to achieve
specified goals with respect to effectiveness, effi-
ciency, and satisfaction in a context of use. Censorship
resistance (U1) describes the probability that a party
can strongly influence the acceptance or refusal of
transactions with a reasonable effort. Little censorship
resistance comes from a low LoD [1]. Therefore, we
operationalize censorship resistance as a probability
dependent on LoD. A high LoD represents high cen-
sorship resistance, which the metric expresses with a
score close to 1.0.
Cost (U2) relates to the implementation and use of
a DLT design. To operationalize cost we draw from a
Total Cost of Ownership (TCO) approach and adjust it
to DLT [38]. Our TCO metric employs the sum of av-
erage Hardware and Electricity (H&E) cost, the aver-
age cost for transactions (e.g., transaction fees), and
the average Maintenance and Servicing (M&S) cost,
all computed per month. Costs for research and human
resources are excluded to keep the calculation feasible
and because they are assumed to be similar for each
DLT design since research and hiring of quality per-
sonal needs to be done for all DLT designs. H&E costs
include the cost for all needed hardware to set up all
necessary components to participate in the distributed
ledger to the extent required by the use case, as well as
the accumulated energy and resource costs of mining
(if mining is performed). M&S costs cover all mainte-
nance costs to keep the network operational and all
services used within or outside the distributed ledger.
4.2. Comparison process for DLT designs
We propose a seven-step process for the holistic
comparison process of DLT designs: use case defini-
tion, requirements definition, DLT design selection,
boundary condition examination, data collection,
analysis, and decision making (cf. Figure 1). In the fol-
lowing, we illustrate the usefulness of the comparison
process for an exemplary use case in IoT. For parsi-
moniousness, we only consider the DLT characteris-
tics consistency (S2) and throughput (P3), which are
consolidated in Table 4. Please refer to Use Cases” in
the supplementary online material for the full compar-
ison (https://bit.ly/2klPK9v).
First, we define the use case and functional re-
quirements for the application during the use case def-
inition. To assess the usefulness of DLT we include
the process of [9] into our comparison process, which
we adapt to DLT in general. The process serves as a
pre-filter to prevent developers from choosing an un-
suitable technology (e.g., DLT vs. centralized data-
base) beforehand. If DLT has been found suitable for
the use case, the process continues.
Second, the pursued values of the individual met-
rics are ascertained as forming criteria of the process
in requirements definition. In the IoT use case, con-
sistency (S2) should be fairly low (< 2 s) because
many IoT networks share and handle real-time sensor
data. IoT networks usually incorporate a large number
of devices and sensors. Therefore, we require through-
put (P3) to be at least 1,000 tps.
Third, the DLT design selection requires to gener-
ate a set of possibly suitable DLT designs for the use
case. A survey on the application of DLT designs in
similar use cases should lead to a set of possible DLT
designs for the following evaluation process. This set
should have a cardinality of at least two DLT designs
but is not restricted by an upper limit. For our exem-
plary use case, we select a set of five DLT designs:
Ethereum, Hyperledger Indy, Tezos, and IOTA, Hash-
graph.
Fourth, during the boundary condition examina-
tion potential compliance, legal, and maturity risks of
members of the previously defined set of possible
DLT designs are considered. For example, newly de-
veloped DLT designs with an error-prone code base
that are hardly ready to use may be examined and then
excluded from the set of candidates. In this step, an
evaluation of previously excluded qualitative charac-
teristics can be made. All DLT designs, which do not
comply with these factors should be excluded from the
set. If too many ledgers are being excluded in this step,
a step back to DLT design selection towards more DLT
design candidates may be necessary. The topics ease
of use, dependencies on third parties (e.g., an enter-
prise, a foundation), auditability restrictions, or data
ownership considerations can be examined to exclude
additional DLT design candidates if they are consid-
ered important for the use case. After reviewing possi-
ble legal, compliance and maturity risks, no hindering
boundary conditions (e.g., prohibiting data-security
legislation, a non-mature code framework) were
found. Therefore, we deem all DLT design candidates
mature and suitable for the exemplary IoT use case.
Fifth, after all factors are included and a final set
of DLT design candidates is generated, we conduct a
data collection to use the developed operationaliza-
tions and to calculate the corresponding metrics. For
our example, we gathered data on transaction latency
and transactions per second for every DLT design out
of the set of DLT design candidates from monitoring
websites (e.g., [39]) and existing studies (e.g., [17,
40]) to calculate the values for consistency (S2) and
throughput (P3).
Sixth, we compose a table to summarize the found
results in the analysis step. The table includes all cri-
teria in the columns and lists all DLT design candi-
dates in the rows. For each DLT design candidate,
every criterion is evaluated on whether the application
requirement is fulfilled by the respective value of a
DLT design candidate. If the calculated value of the
metric fulfills the particular application requirement,
we rate it 1. If a criterion is only partially fulfilled or
the application requirement is very close to the actual
value of the DLT design, we rate it 0.5. Otherwise, the
table entry is rated 0.
Finally, the decision making is performed by sum-
ming up the ratings for each DLT design candidate,
leading to a total score. The DLT design with the high-
est score represents the assumed best suitable DLT de-
sign for the application on DLT. In the exemplary IoT
use case, Hyperledger Indy scored best (cf. Table 4).
Figure 1. Sequential steps of the process to compare DLT designs
Table 4. Evaluation of consistency and
throughput inside the process to compare
DLT designs
Consistency
(< 2 s)
Throughput
(> 1000 tps)
DLT Design Candidates
Value [s]
Score
Value
[tps]
Score
Ethereum
<15
0
20
0
0
Hashgraph
n.A.2
0
> 10,000
1
1
Hyperledger Indy
<21
1
>3,500
1
2
Tezos
<60
0
40
0
0
IOTA
n.A.2
0
500-800
0.5
0.5
1: Transaction latency is taken from a Hyperledger Fabric eval-
uation. Since both DLT designs employ a similar PBFT consen-
sus algorithm with low latencies the data is comparable.
2: Transaction confirmation is non-deterministic
DLT De sign
Selection
Data
Collection
Boundary
Condition
Exam ination
Analysis Decision
Making
Use Case
Definition Requirements
Definition Enough DLT can didates?
Yes
No
Is DLT useful?
No
Yes
5. Discussion
In this work, we present a comprehensive overview
of DLT characteristics and their corresponding opera-
tionalization that can be used as criteria for a holistic
comparison. Furthermore, we generated a first ap-
proach for a structured comparison process including
research from DLT, distributed databases, and com-
puter science. The validation of the generated process
for comparison indicates its validity because the cal-
culated suitability ranking is coherent with the in-
tended purposes of the included DLT design candi-
dates. For each use case, the process proposed a suita-
ble DLT design for which proof of concepts in the re-
spective fields have already been developed.
We regard the proposed operationalization as a
first approach for the operationalization of DLT de-
signs for a holistic comparison. Due to the strong focus
of the included criteria for performance and security,
the process will probably perform sufficiently for DLT
designs that are designed for a similar purpose.
5.1. Implications
The presented operationalization of DLT charac-
teristics supports a better comparison of DLT designs
and helps to quantify their advantages. Using the op-
erationalizations, the presented process supports the
selection of a suitable DLT design for applications on
DLT. The process facilitates and structures decision
making for choosing a DLT design, which avoids un-
necessary overhead and improves decision making.
The developed process synthesizes extant research
on DLT from research in computer science as we eval-
uated different operationalizations for DLT character-
istics and present a set of applicable operationaliza-
tions for selected criteria. Through the selection of op-
erationalizations applicable to DLT characteristics, we
support research on the identification of a suitable
DLT design in information systems. The presented op-
erationalizations of DLT characteristics support the
quantification of dependencies between DLT charac-
teristics, which broadens the scope for a comprehen-
sive analysis of DLT designs [33] and helps to reveal
the weaknesses of current DLT designs.
5.2. Limitations and future research
As the selection and operationalization of DLT
characteristics are based primarily on literature and the
presented process has only been evaluated in three use
cases, we cannot generalize the presented process
without limitation. As not all DLT characteristics have
a corresponding characteristic in research on database
or distributed systems monitoring (e.g., stale block
rate), we developed own operationalizations. How-
ever, these measures have not been evaluated in the
field, yet. Some operationalization concepts are re-
stricted to or have a higher significance with block-
chain-based DLT designs. During the data collection
for the validation of the process, it became obvious
that DLT and especially some of the chosen DLT de-
signs form a fairly new research topic, thus discover-
ing a need for additional research and practical meas-
urement of DLT characteristics. This work relies on
previous research on DLT characteristics [1, 24].
Thus, we also used the DLT property usability accord-
ing to the examined literature.
In future research, the presented comparison pro-
cess should be applied and evaluated in the field to as-
sess its validity and overcome potential challenges in
its usefulness. Additionally, the scoring model and the
use of weights for the respective importance of DLT
characteristics should be investigated. Since several
DLT characteristics are not yet operationalizable, it is
of high interest to generate new operationalizations for
such DLT characteristics to obtain a holistic view of
DLT designs and their benefits and potential con-
straints. Since the importance of cross-chain technol-
ogy in DLT increases [25], additional metrics and op-
erationalizations should be investigated and developed
in order to make different cross-chain technology ap-
proaches comparable with each other.
6. References
[1] Kannengießer, N., S. Lins, T. Dehling, and A. Sunyaev,
“What Does Not Fit Can be Made to Fit! Trade-Offs in Dis-
tributed Ledger Technology Designs”, 52nd Hawaii Inter-
national Conference on System Sciences, (2019).
[2] O’Donoghue, O., A.A. Vazirani, D. Brindley, and E.
Meinert, “Design Choices and Trade-Offs in Health Care
Blockchain Implementations: Systematic Review”, Journal
of Medical Internet Research 21(5), 2019, pp. e12426.
[3] Nelson, J., M. Ali, R. Shea, and M.J. Freedman, “Extend-
ing Existing Blockchains with Virtualchain”, 2016.
https://www.zurich.ibm.com/dccl/papers/nelson_dccl.pdf
[4] Pass, R., and E. Shi, “The Sleepy Model of Consensus”,
Advances in Cryptology ASIACRYPT 2017, (2017), 380
409.
[5] Eyal, I., and E.G. Sirer, “Majority is not Enough: Bitcoin
Mining is Vulnerable”, In Financial Cryptography and Data
Security. 2014.
[6] Weber, I., V. Gramoli, A. Ponomarev, et al., “On Avail-
ability for Blockchain-Based Systems”, 2017 IEEE 36th
Symposium on Reliable Distributed Systems, (2017), 6473.
[7] Gervais, A., G.O. Karame, K. Wüst, V. Glykantzis, H.
Ritzdorf, and S. Capkun, “On the Security and Performance
of Proof of Work Blockchains”, Proceedings of the 2016
ACM SIGSAC Conference on Computer and Communica-
tions Security, (2016), 316.
[8] Hyperledger Performance and Scale Working Group,
“Hyperledger Blockchain Performance Metrics”, 2018.
https://www.hyperledger.org/wp-content/up-
loads/2018/10/HL_Whitepaper_Metrics_PDF_V1.01.pdf
[9] Lo, S.K., X. Xu, Y.K. Chiam, and Q. Lu, “Evaluating
Suitability of Applying Blockchain”, 2017 22nd Interna-
tional Conference on Engineering of Complex Computer
Systems, (2017), 158161.
[10] Wüst, K., and A. Gervais, “Do you need a Block-
chain?”, 2017. https://eprint.iacr.org/2017/375.pdf
[11] Zhang, K., and H.-A. Jacobsen, “Towards Dependable,
Scalable, and Pervasive Distributed Ledgers with Block-
chains”, 2018 IEEE 38th International Conference on Dis-
tributed Computing Systems, (2018), 13371346.
[12] Lamport, L., R. Shostak, and M. Pease, “The Byzantine
Generals Problem”, ACM Transactions on Programming
Languages and Systems 4(3), 1982, pp. 382401.
[13] Natoli, C., and V. Gramoli, “The Blockchain Anom-
aly”, 2016 IEEE 15th International Symposium on Network
Computing and Applications, (2016), 310317.
[14] Bermbach, D., E. Wittern, and S. Tai, Cloud Service
Benchmarking: Measuring Quality of Cloud Services from a
Client Perspective, 2017.
[15] Decker, C., and R. Wattenhofer, “Information propaga-
tion in the Bitcoin network”, IEEE P2P 2013 Proceedings,
(2013), 110.
[16] Bott, J., and U. Milkau, “Towards a framework for the
evaluation and design of distributed ledger technologies in
banking and payments.”, Journal of Payments Strategy &
Systems 10(2), 2016, pp. 153171.
[17] Dinh, T.T.A., J. Wang, G. Chen, R. Liu, B.C. Ooi, and
K.-L. Tan, “BLOCKBENCH: A Framework for Analyzing
Private Blockchains”, ACM International Conference on
Management of Data, (2017), 10851100.
[18] Konheim, A.G., Hashing in computer science: fifty
years of slicing and dicing, John Wiley & Sons, Hoboken,
N.J, 2010.
[19] Glaser, F., and L. Bezzenberger, “Beyond Cryptocur-
rencies - A Taxonomy of Decentralized Consensus Sys-
tems”, 23rd European Conference on Information Systems,
(2015), 118.
[20] Gray, J., ed., The Benchmark handbook: for database
and transaction processing systems, M. Kaufmann Publish-
ers, San Mateo, Calif, 1991.
[21] Huppler, K., “The Art of Building a Good Benchmark”,
Performance Evaluation and Benchmarking, (2009), 1830.
[22] Hyperledger Foundation, “Hyperledger Caliper”, 2018.
https://github.com/hyperledger/caliper
[23] Kaner, C., S. Member, and W.P. Bond, “Software En-
gineering Metrics: What Do They Measure and How Do We
Know?”, In METRICS 2004. IEEE CS, (2004).
[24] Kannengießer, N., S. Lins, T. Dehling, and A. Sunyaev,
“Mind the Gap: Trade-Offs between Distributed Ledger
Technology Characteristics”, arXiv:1906.00861 [cs], 2019.
[25] Kannengießer, N., M. Pfister, M. Greulich, S. Lins, and
A. Sunyaev, “Bridges Between Islands: Cross-Chain Tech-
nology for Distributed Ledger Technology”, 53rd Hawaii
International Conference on System Sciences, (2020).
[26] Yli-Huumo, J., D. Ko, S. Choi, S. Park, and K. Smo-
lander, “Where Is Current Research on Blockchain Technol-
ogy?”, PLOS ONE 11(10), 2016, pp. e0163477.
[27] Göbel, J., and A.E. Krzesinski, “Increased block size
and Bitcoin blockchain dynamics”, 2017 27th International
Telecommunication Networks and Applications Conference,
(2017), 16.
[28] Ahmed, P.K., and M. Rafiq, “Integrated benchmarking:
a holistic examination of select techniques for benchmarking
analysis”, Benchmarking for Quality Management & Tech-
nology 5(3), 1998, pp. 225242.
[29] Doll, W.J., X. Deng, and J.A. Scazzero, “A process for
post-implementation IT benchmarking”, Information &
Management 41(2), 2003, pp. 199212.
[30] Hileman, G., and M. Rauchs, “2017 Global Blockchain
Benchmarking Study”, SSRN Electronic Journal, 2017.
[31] Deloitte, “Breaking blockchain open”, 2018.
https://www2.deloitte.com/content/dam/Deloitte/cz/Docu-
ments/financial-services/cz-2018-deloitte-global-block-
chain-survey.pdf
[32] Liu, Y., X. Liu, C. Tang, J. Wang, and L. Zhang, “Un-
linkable Coin Mixing Scheme For Transaction Privacy En-
hancement of Bitcoin”, IEEE Access 6, 2018, pp. 11.
[33] Anceaume, E., T. Lajoie-Mazenc, R. Ludinard, and B.
Sericola, “Safety analysis of Bitcoin improvement pro-
posals”, 2016 IEEE 15th International Symposium on Net-
work Computing and Applications, (2016), 318325.
[34] Jogalekar, P.P., and C.M. Woodside, “A scalability
metric for distributed computing applications in telecommu-
nications”, In Teletraffic Contributions for the Information
Age. 1997, 101110.
[35] McClusky, E.J., and S. Mitra, “Fault Tolerance”, In
Computer Science Handbook. Chapman & Hall/CRC, Boca
Raton, Fla, 2004.
[36] Haerder, T., and A. Reuter, “Principles of transaction-
oriented database recovery”, ACM Computing Surveys
15(4), 1983, pp. 287317.
[37] Goldwasser, S., and M. Bellare, Lecture Notes on Cryp-
tography, Cambridge, Massachusetts, USA, 2008.
[38] Ellram, L.M., “Total cost of ownership: an analysis ap-
proach for purchasing”, International Journal of Physical
Distribution & Logistics Management 25(8), 1995, pp. 423.
[39] Etherscan, “Etherscan”, https://etherscan.io/
[40] Nasir, Q., I.A. Qasse, M. Abu Talib, and A.B. Nassif,
“Performance Analysis of Hyperledger Fabric Platforms”,
Security and Communication Networks 2018, 2018, pp. 1
14
... DLT enables unprecedented traceability and efficient data decentralisation by associating linked records of cryptographically secured, propagated and validated across peers and nodes (Red 2017;Rückeshäuser 2017;Seok, Park, and Park 2019). DLT encompasses multiple areas of computer science (i.e., information systems, cryptography and databases) (Poels et al. 2019;Gräbe et al. 2020); however, this research is more aligned with the information system aspects of DLT. Presently, prior studies are mostly focused on the technological aspects and challenges of adopting DLT for peer to peer (P2P) enterprise operations (Ribitzky et al. 2018). ...
... Distributed ledgers are replicated, and immutable synchronised data structures (or databases) shared and managed by the equally untrusted network of peers or nodes (Gräbe et al. 2020). Commonly, DLT platforms comprise first of peer-to-peer network for network maintenance and peer interaction with the aim of sharing distributed ledger transactions amongst all peers within the network ). ...
... This comprises DLT characteristics regarding the achievement of a given task assessed against standards of completeness, accuracy and speed (scalability, confirmation latency and throughput) (Gräbe et al. 2020) Security ...
Article
Full-text available
Digital transformation of Collaborative Enterprise (CE) both in terms of planning and implementation, relies on new business models and innovative technologies. One of such technologies is Distributed Ledger Technology (DLT) which deploys a decentralized approach requiring no need for intermediaries, thereby facilitating decentralized enterprise operations. Accordingly, this article provides a systematic review and theoretical background of the current state and applications of DLTs within CEs by exploring how this disruptive technology can digitally transform CE practices. Grounded on the review of existing literature a comprehensive characteristics, application, governance, stakeholders, and emerging security threats faced during DLT adoption and mitigation strategies are presented. More importantly, a model is proposed based on the factors that inhibit DLT adoption in CE. Findings from this study provide implications to support smart digital transition in CE for smart enterprise operations.
... Extant research has already presented various approaches to compare distributed ledgers from a technical perspective (e.g., considering availability or throughput) [11], [12] or from an organizational perspective (e.g., trust in other stakeholders) [13], [14]. These studies focus on comparing distributed ledgers' system behaviors in a static manner, for instance, by comparing the maximum throughput of different distributed ledgers in order to asses their suitability for a specific DLT application [11], [12]. ...
... Extant research has already presented various approaches to compare distributed ledgers from a technical perspective (e.g., considering availability or throughput) [11], [12] or from an organizational perspective (e.g., trust in other stakeholders) [13], [14]. These studies focus on comparing distributed ledgers' system behaviors in a static manner, for instance, by comparing the maximum throughput of different distributed ledgers in order to asses their suitability for a specific DLT application [11], [12]. However, in order to gain insights into the system behavior of distributed ledgers, either exhaustive field tests or controlled simulations that allow understanding the complex dependencies between DLT characteristics are required. ...
... The most direct approach for assessing the behavior of a distributed ledger is to test a DLT application on a running system with real users. Since changes in a distributed ledger's configuration and other invasive interventions may disrupt its regular operation, evaluations of live systems are usually limited to black-box approaches like benchmarking (e.g., [11], [12]). ...
Conference Paper
Full-text available
Interest in applications based on distributed ledger technology (DLT) is on the rise, with corporations worldwide shifting from simply exploring DLT's potential to creating productive business cases. However, despite the existence of numerous DLT applications, developers are still lacking proper tools and instruments for evaluating system behavior (e.g., performance) of their applications on different distributed ledgers before deployment. Since the behavior of such applications is highly dependent on the characteristics of the distributed ledger they are built upon and changing the distributed ledger after deployment of an application is difficult, selecting the wrong ledger can have severe negative consequences. To address the issue, we conducted an extensive literature review to identify and synthesize published modeling and simulation approaches for distributed ledgers. Based on the results, this paper also presents a research agenda to improve modeling of the system behavior of distributed ledgers and to support the development of distributed ledgers and viable DLT applications. In doing so, we facilitate informed decision-making for suitable modeling or simulation approach, which helps application developers to identify a suitable distributed ledger before implementing an application. In addition, our work contributes to science, as we provide a comprehensive overview and analysis of extant simulation and modeling approaches in DLT that accumulates the current state-of-the-art in the rapidly growing DLT field.
... The second trade-off is known as the "blockchain trilemma", illustrated in Fig. 3, which is essentially the reformulation of the PACELC theorem for the blockchain domain [20]. In particular, the trilemma illustrates the conjecture that a blockchain system cannot exhibit maximum decentralization, security and scalability (performance) at the same time. ...
... This is fundamentally the strategy adopted with permissioned ledgers, in which a central registrar is introduced to authenticate users, but in this case decentralization is traded for a performance gain. Again, a tradeoff must be chosen, as the trilemma warns us that no consensus protocol can ensure full security, decentralization and scalability at the same time [20]. ...
... In addition, the patient cannot track his/her personal data access rights. [48] Overcoming barriers between network environments Scalability IoT, DA, GA [11], [38], [79] An increase in the load on the centralized node occurs with an increase in traffic or the number of users All DLT methods, DAG, vDLT, Sharding, Sidechain, and cross-chain Creation of applications, where data exchange is based on the distributed registry principles Identity, security, and data management IoT, IR, GA [28], [45] The identification of an organization's assets (including people, buildings, cars, systems, and information assets) DLT design in system functionality GA [52], [30] Depending on the chosen design, the system may have certain functions. ...
... The article [30] outlines DLT characteristics that are selected according to an in-depth comparison of DLT elements. In addition, a benchmarking process is proposed to structure of the DLT projects according to the application requirements. ...
Article
Full-text available
Distributed Ledger Technology (DLT) is making the first steps toward becoming a solution for the growing number of various decentralized systems worldwide. Unlike pure Blockchain, DLT finds many uses across different industries, including eHealth, finance, supply chain monitoring, and the Internet of Things (IoT). One of the vital DLT features is the ability to provide an immutable and commonly verifiable ledger for larger-scale and highly complex systems. Today’s centralized systems can no longer guarantee the required level of availability and reliability due to the growing number of the involved nodes, complicated heterogeneous architectures, and task load, while the publicly available distributed systems are still in their infancy. This paper aims to provide an exhaustive topical review of the state-of-the-art of Distributed Ledger Technology applicability in various sectors. It outlines the importance of the practical integration of technology-related challenges, as well as potential solutions.
... Limited by the trilemma, an IoT developer willing to improve the network scalability may chose a consensus protocol less expensive than PoW or reduce the mining difficulty to 3 Recall that the Bitcoin protocol enforces consistency introducing, by design, an average latency of 10 minutes per block, and further recommends to consider a block as "unconfirmed" until it becomes 6 blocks deep (Section 2.2). Figure 6: The blockchain trilemma is illustrated by a triangle like the one in this figure where one cannot draw a single point that is close to all the triangle corners, meaning that a trade-off among the three properties must be chosen [80]. ...
... This is fundamentally the strategy adopted with permissioned ledgers, in which a central registrar is introduced to authenticate users, but in this case decentralization is traded for a performance gain. Again, a trade-off must be chosen, as the trilemma warns us that no consensus protocol can ensure full security, decentralization and scalability [80]. An IoT developer should therefore choose a consensus protocol and a blockchain-based system only after having clearly identified the application requirements, choosing the most appropriate trade-off. ...
Preprint
Full-text available
The use of the term blockchain is documented for disparate projects, from cryptocurrencies to applications for the Internet of Things (IoT), and many more. The concept of blockchain appears therefore blurred, as it is hard to believe that the same technology can empower applications that have extremely different requirements and exhibit dissimilar performance and security. This position paper elaborates on the theory of distributed systems to advance a clear definition of blockchain that allows us to clarify its role in the IoT. This definition inextricably binds together three elements that, as a whole, provide the blockchain with those unique features that distinguish it from other distributed ledger technologies: immutability, transparency and anonimity. We note however that immutability comes at the expense of remarkable resource consumption, transparency demands no confidentiality and anonymity prevents user identification and registration. This is in stark contrast to the requirements of most IoT applications that are made up of resource constrained devices, whose data need to be kept confidential and users to be clearly known. Building on the proposed definition, we derive new guidelines for selecting the proper distributed ledger technology depending on application requirements and trust models, identifying common pitfalls leading to improper applications of the blockchain. We finally indicate a feasible role of the blockchain for the IoT: myriads of local, IoT transactions can be aggregated off-chain and then be successfully recorded on an external blockchain as a means of public accountability when required.
... Therefore, blockchain technologies in platform models [14] represent a unique opportunity to manage new enriched participatory processes organized by public institutions in a more effective and efficient manner [15]. Thus, the objective of this paper is to propose an e-Participation model that uses DLTs systems [16] [17] to implement a new tool to facilitate and empower citizen participation. The model integrates the opportunities offered by DLTs systems [18] in relation to public governance processes in order to develop new tools for citizen empowerment. ...
Article
Currently, Distributed Ledger Technologies (DLTs) and, especially, Blockchain technology represent a great opportunity for public institutions to improve citizen participation and foster democratic innovation. These technologies facilitate the simplification of processes and provide secure management of recorded data, guaranteeing the transmission and public transparency of information. Based on the combination of a Blockchain as a Service (BaaS) platform and G-Cloud solutions, our proposal consists of the design of an e-Participation model that uses a tokenizable system of the actions and processes undertaken by citizens in participatory processes providing incentives to promote greater participation in public affairs. In order to develop a sustainable, scalable and resilient e-Participation system, a new blockchain concept, which organizes the blocks as a neural system, is combined with the implementation of a virtual token to reward participants. Furthermore, this virtual token is deployed through a smart contract that the block itself produces, containing information about the transaction and all the documents involved in the process. Finally, our Neural Distributed Ledger (NDL) framework facilitates the interconnection of blockchain networks in a transparent, certified, secure, auditable, scalable and traceable way.
... However, the presented processes are highly abstract and generic, they focus on questions related to whether a distributed ledger is useful at all. Some articles consider select DLT designs and compare them but hardly address causes for the viability of investigated DLT designs for applications (e.g., [17,76,77]). Other articles address the degree of decentralization (e.g., [68,70]). ...
Article
Full-text available
When developing peer-to-peer applications on distributed ledger technology (DLT), a crucial decision is the selection of a suitable DLT design (e.g., Ethereum), because it is hard to change the underlying DLT design post hoc. To facilitate the selection of suitable DLT designs, we review DLT characteristics and identify trade-offs between them. Furthermore, we assess how DLT designs account for these trade-offs and we develop archetypes for DLT designs that cater to specific requirements of applications on DLT. The main purpose of our article is to introduce scientific and practical audiences to the intricacies of DLT designs and to support development of viable applications on DLT.
Article
Full-text available
Blockchain is increasingly used for registering, authenticating and validating digital assets (financial assets, real estate, etc.) and transactions, governing interactions, recording data and managing identification among multiple parties in a trusted, decentralized, and secure manner. Today, a large variety of blockchain technologies is expanding in order to fulfill technical and non-technical needs and requirements. Within this context, determining and most importantly evaluating the characteristics/performance of a given blockchain platform is crucial for system designers before deploying it. A number of blockchain simulators have been proposed in the literature over the past few years, as reviewed in this paper, but are often limited in several respects (lack of extensibility, do not allow for evaluating all aspects of a blockchain. . .). This paper extends and improves a state-of-the-art simulator (BlockSim) into a new simulator called “BlockPerf" to overcome those limitations. Both simulators are compared based on a real-life (benchmarking) bitcoin scenario, whose results show that BlockPerf provides more realistic results than BlockSim, improving by around ≈50% (in average) the outcomes.
Article
Full-text available
Introduction: Distributed ledger technology (DLT) is one of the most significant fields covering various aspects of modern ICT systems. Centralized systems of today can no longer guarantee the required level of availability and reliability, while broadly available distributed ones are still in the infancy. Purpose: Analysis of the applicability of DLT to various industries such as economics, energy, finance, logistics, and the Internet of Things. Results: The article outlines the main challenges of the DLT technology integration, such as the lack of a unified system for data storage, the need to ensure an appropriate level of data confidentiality, integration into the existing competency systems, the issues of а distributed system interaction with resource-constrained IoT devices, the lack of proper management tools for distributed systems, and network scalability. The main contribution of this paper is a systematic overview of the integration challenges followed by potential solutions and future perspectives.
Conference Paper
Full-text available
Since blockchain's emergence in 2008, we see a kaleidoscopic variety of applications built on distributed ledger technology (DLT) today, including applications for financial services, healthcare, or the Internet of Things. Each application comes with specific requirements for DLT characteristics (e.g., high throughput, scalability). However, trade-offs between DLT characteristics restrict the development of a DLT design (e.g., Ethereum, blockchain) that fits all use cases' requirements. Separated DLT designs emerged, each specialized to suite dedicated application requirements. To enable the development of more powerful applications on DLT, such DLT islands must be bridged. However, knowledge of cross-chain technology (CCT) is scattered across scientific and practical sources. Therefore, we examine this diverse body of knowledge and provide comprehensive insights into CCT by synthesizing its underlying characteristics, evolving patterns, and use cases. Our findings resolve contradictions in the literature and provide avenues for future research in an emerging scientific field.
Article
Full-text available
Background: A blockchain is a list of records that uses cryptography to make stored data immutable; their use has recently been proposed for electronic medical record (EMR) systems. This paper details a systematic review of trade-offs in blockchain technologies that are relevant to EMRs. Trade-offs are defined as "a compromise between two desirable but incompatible features." Objective: This review's primary research question was: "What are the trade-offs involved in different blockchain designs that are relevant to the creation of blockchain-based electronic medical records systems?" Methods: Seven databases were systematically searched for relevant articles using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). Papers published from January 1, 2017 to June 15, 2018 were selected. Quality assessments of papers were performed using the Risk Of Bias In Non-randomized Studies-of Interventions (ROBINS-I) tool and the Critical Assessment Skills Programme (CASP) tool. Database searches identified 2885 articles, of which 15 were ultimately included for analysis. Results: A total of 17 trade-offs were identified impacting the design, development, and implementation of blockchain systems; these trade-offs are organized into themes, including business, application, data, and technology architecture. Conclusions: The key findings concluded the following: (1) multiple trade-offs can be managed adaptively to improve EMR utility; (2) multiple trade-offs involve improving the security of blockchain systems at the cost of other features, meaning EMR efficacy highly depends on data protection standards; and (3) multiple trade-offs result in improved blockchain scalability. Consideration of these trade-offs will be important to the specific environment in which electronic medical records are being developed. This review also uses its findings to suggest useful design choices for a hypothetical National Health Service blockchain. International registered report identifier (irrid): RR2-10.2196/10994.
Conference Paper
Full-text available
Distributed ledger technology (DLT), including blockchain, enables secure processing of transactions between untrustworthy parties in a decentralized system. However, DLT is available in different designs that exhibit diverse characteristics. Moreover , DLT characteristics have complementary and conflicting interdependencies. Hence, there will never be an ideal DLT design for all DLT use cases; instead, DLT implementations need to be configured to contextual requirements. Successful DLT configuration requires, however, a sound understanding of DLT characteristics and their interdependencies. In this manuscript, we review DLT characteristics and organize them into six groups. Furthermore, we condense interdependencies of DLT characteristics into trade-offs that should be considered for successful deployment of DLT. Finally, we consolidate our findings into DLT archetypes for common design objectives , such as security, usability, or performance. Our work makes extant DLT research more transparent and fosters understanding of interdependencies and trade-offs between DLT characteristics.
Article
Full-text available
Blockchain is a key technology that has the potential to decentralize the way we store, share, and manage information and data. One of the more recent blockchain platforms that has emerged is Hyperledger Fabric, an open source, permissioned blockchain that was introduced by IBM, first as Hyperledger Fabric v0.6, and then more recently, in 2017, IBM released Hyperledger Fabric v1.0. Although there are many blockchain platforms, there is no clear methodology for evaluating and assessing the different blockchain platforms in terms of their various aspects, such as performance, security, and scalability. In addition, the new version of Hyperledger Fabric was never evaluated against any other blockchain platform. In this paper, we will first conduct a performance analysis of the two versions of Hyperledger Fabric, v0.6 and v1.0. The performance evaluation of the two platforms will be assessed in terms of execution time, latency, and throughput, by varying the workload in each platform up to 10,000 transactions. Second, we will analyze the scalability of the two platforms by varying the number of nodes up to 20 nodes in each platform. Overall, the performance analysis results across all evaluation metrics, scalability, throughput, execution time, and latency, demonstrate that Hyperledger Fabric v1.0 consistently outperforms Hyperledger Fabric v0.6. However, Hyperledger Fabric v1.0 platform performance did not reach the performance level in current traditional database systems under high workload scenarios.
Article
Bitcoin combines a peer-to-peer network and cryptographic algorithm to implement a distributed digital currency system, which keeps all transaction history on a public blockchain. Since all transactions recorded on the blockchain are public to everyone, Bitcoin users face a threat of leaking financial privacy. Many analysis and deanonymization approaches have been proposed to link transaction records to real identities. To eliminate this threat, we present an unlinkable coin mixing scheme that allows users to mix their bitcoins without trusting a third party. This mixing scheme employs a primitive known as ring signature with elliptic curve digital signature algorithm(ECDSA) to conceal the transfer of coins between addresses. The mixing server is only able to check whether the output addresses belong to its customers, but it can not tell which address owned by which customer. Customers do not have to rely on the reputation of a third party to ensure his money will be returned and his privacy will not be leaked. This scheme needs no modifications on current Bitcoin system and is convenient to deploy by any communities. We implemented a prototype of our scheme and tested it under the Bitcoin core’s regtest mode. Security and privacy of our mixing scheme are ensured through the standard ring signature and ECDSA unforgeability.