Research paper

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/337527299

Blockchain-Based Electronic Voting System for Elections in Turkey

Conference Paper · September 2019

DOI: 10.1109/UBMK.2019.8907102

CITATION

1 READS

752

4 authors, including:

Some of the authors of this publication are also working on these related projects:

Attack Detection And Countermeasures Simulation ADAX View project

Alperen Kantarcı

Istanbul Technical University

5 PUBLICATIONS   1 CITATION   

SEE PROFILE

Serif Bahtiyar

Istanbul Technical University

39 PUBLICATIONS   112 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Alperen Kantarcı on 26 November 2019.

The user has requested enhancement of the downloaded file.

Blockchain-Based Electronic Voting System for

Elections in Turkey

Rumeysa Bulut, Alperen Kantarcı, Safa Keskin, Şerif Bahtiyar Faculty of Computer and Informatics

Istanbul Technical University

Istanbul, Turkey

{bulutr, kantarcia, keskinsaf, bahtiyars}@itu.edu.tr

Abstract—Traditional elections satisfy neither citizens nor

political authorities in recent years. They are not fully secure

since it is easy to attack votes. It threatens also privacy and

transparency of voters. Additionally, it takes too much time to

count the votes. This paper proposes a solution using Blockchain

to eliminate all disadvantages of conventional elections. Security

and data integrity of votes is absolutely provided theoretically.

Voter privacy is another requirement that is ensured in the

system. Lastly, waiting time for results decreased significantly in

proposed Blockchain voting system.

Keywords—blockchain, e-voting, electronic voting, internet

voting

I. INTRODUCTION

Elections are the organizations that is supposed to bring

democracy into countries. They mostly play a crucial role in

the future of a country and citizens life. Therefore, it has much

importance for every single person involved in these

elections. Regardless of the organization, elections have to be

trustworthy in its nature. They have to ensure people’s

privacy and vote’s security. Additionally, the authority which

is responsible for counting votes should not spend too much

time on counting votes since waiting long period of time for

results increases concerns about manipulation of results.

However, due to the different reasons depending on the areas

that elections have been made, trust has been a controversial

issue for each election. Especially, paper elections are

managed by a centralized authority, there is always a risk to

manipulate ballots and election results [1]. For example, there

are disagreements about security and privacy of elections and

whether all votes are counted correctly in Turkey. Moreover,

time factor can be a challenging issue for announcing results.

The speed of counting votes and making public the unofficial

results takes seven to eight hours in Turkey, whereas the

official results are announced in ten days approximately. The

official results of 24th June General Elections of Turkey were

expounded after eleven days on 5th July.

There are some attempts to remove problems of

traditional election system. These attempts try to benefit from

online systems to automate the whole process. Electronic

voting was used in elections of Austrian Federation of

Students in 2009 [2] and in some elections in Switzerland [3].

Although e-voting makes selection operation easy, privacy

and security worries still continue. To dissipate problems of

both conventional and e-voting elections, e-voting can be

improved using Blockchain mechanism. Blockchain has

impressive features to overcome troubles of voter’s security,

privacy and data integrity of votes.

Blockchain is an inalterable and an easy confirmable

system [4]. Under favor of these qualifications, Blockchain

has a significant potential to be an alternative to traditional

elections. It brings smart solutions to central authority

problem in terms of all blocks having all data in the chain.

Also, it is impossible to change an information in a block

since it is discerned by other blocks which have whole data.

Consequently, Blockchain increases the security of

information by keeping the entire data in all blocks, and

removes the need for an official center to provide a secure

election [5]. As mentioned before, counting votes and making

election results publicly available takes considerable time.

Blockchain solves this problem by its nature. Since the last

node on the chain keeps all information, it is enough to look

for only the last node for the results. This reduces the waiting

time dramatically. Thus, incomplete and official results are

explained at the same time.

Since each country has different laws and

implementations, proposing a definitive structure is almost

impossible. The suggested solution in this paper is

specifically for problems of conventional paper elections in

Turkey. Despite the solution is specific to one country, it may

be taken as a general application, and can be customized to

other countries. Security and privacy of votes and voters and

the speed of counting votes and announcing the results are

discussed in the solution. A representative model of the

system is presented in Proposed Solution section.

The paper is organized as follows. In Section II, related

works about e-voting and blockchain are discussed. In

Section III, general architecture of the proposed blockchain

based e-voting election system is explained and modeled with

supporting materials. At the following section, Section 4,

proposed system is analyzed from different aspects. In the

final section, Section 5, future work on the system is

discussed and conclusions about the research have been

made.

II. E-VOTING SYSTEMS AND BLOCKCHAIN

E-voting in different areas has been considered for a while

by some countries. The pioneer country in e-voting process is

Estonia which held online voting between 2005 and 2007 [6].

On the other hand, Blockchain-based election has not been

commonly applied yet. It is in development process in recent

years. South Korea is a noticeable specimen which brought

Blockchain-based election to a successful conclusion in 2017

[7]. Additionally, there are some papers that widen viewpoint

about Blockchain-based voting by offering different ideas.

However, not all of them is found useful by authors of this

paper.

In [8], the voting process relies on citizen’s email address

that can be hacked or manipulated easily. To be obvious, there

will be always some people who registers to the system using

someone else’s mail address and votes on behalf of them. For

example, a grandson may open an email address for his

grandparents from different devices, and cast their votes. This

method guarantees none of the required qualifications such as

security, data integrity or privacy that an e-voting system has

system. For such a system, stealing votes or changing votes

are totally.

Researchers of the [9], proposed a peer-to-peer blockchain

based voting system. Main focus of this research is to protect

the anonymity of the ballots and commitment of the votes to

the blockchain. According to this purpose, they propose a

unique vote commitment format. Their solution has solid base

for such a vote commitment format but we propose a different

system that leans on another system that is maintained by

government. In this purpose we preferred to use a structure

for chains that consists of different key-value pairs represents

vote itself.

Another paper proposes a solution with a database

alongside Blockchain [1]. The authors designed a system that

creates blocks following collection of ballots from voters to

keep them in a database until the end of election process [1].

In this paper, authors tried to eliminate the need for a

database.

Blockchain is being used for various areas such as IoT.

Since there are so many different devices and each of them

are processing different data, new approaches emerged from

this area. In [10] author discusses the blockchain IoT

interactions in the paper and one of the model that have been

discussed is hybrid model that uses different chains in

different layers and levels which inspired us in our blockchain

based e-voting system.

Another work in the related area propose a one-time ring

signature in order to ensure the anonymity of the voting

citizen [11]. However, each candidate in the election needs

public key pair in this architecture and adding a new

candidate increase the complexity of signing process and at

every node demanded CPU power increases. One time ring

signature architecture does not depend any trusting center but

in our special case we give the authority of the selecting a

candidate to the government which is the trusting center for

the election.

Previous researches are reported according to the most

remarkable properties in Table 1.

III. BLOCKCHAIN-BASED VOTING

Considering today’s technology, blockchain may create

one of the most prominent alternative to traditional voting in

terms of security, consistency and speed. While designing a

chain for voting in a crowded country, the system should be

secure. Many aspects should be considered in order to

construct a secure blockchain-based election system. First

factor is human for such a system. In the solution, human

interference is absolutely prohibited.

The proposed system will be consisting of nodes

(computers in design) that is closed to human interference.

Any input that cannot be considered as vote will be ignored

in this system. For such a system, stealing votes or changing

votes are totally blocked. Second issue is saving system from

hackers. In order to manipulate votes, hackers need to enter

the system as a citizen at proposed solution. Also, it is

guaranteed that a citizen can only vote for one time. When

citizen cast a ballot, e-government system will be informed

without revealing any information about vote. Then, e-

government system marks that person as voted. Since the

system takes electorate data from e-government, it is not

possible for a marked person to vote again. Although a hacker

is obtained the citizen information and entered to the system,

he cannot vote more than one time.

In a blockchain system, every transaction is related to the

previous one. So, changing an accepted transaction is

impossible for such a system. Due to the consistency of the

blockchain, data will always be consistent and voting will be

reliable. In a case of manipulation of the system such as

changing votes or stealing votes, other connected nodes will

already be synchronized. So, the changed data will be

identified instantly. Details of the system will be explained

below after the use case diagram and explanation of it.

As you can see in the Fig. 1, standard use case of the

election system is about the citizen and government.

Government in this system only provides the authorization of

Researches Architecture and Design Security Considerations User Authentication

[1] ● Blockchain permission ● Maintain data integrity

● Data integrity ● Digital Signature

● Asymmetric authentication

[4]

● Smart contract ● POA Permissioned

blockchain

● Exonum, Quorum and Geth Frameworks

● Secure authentication via identify verification

● Does not allow to trace voters from votes

● Transparent

● Identity verification service

[6]

● Multi-tiered ● Two Factor

Authentication

● Encryption based on public-private keys

● User authentication ● Monitoring and auditing for

data integrity

● Risk of voter to forget their ID, password

● Randomly generated password to use on polling station

[8]

● Zcash tokens ● Authentication with

Challenge-Handshake

Authentication Protocol

● Anonymity ● Privacy ● Transparency

● Challenge-Handshake Authentication Protocol (CHAP)

● The voters’ email addresses

[11] ● Smart contract on

Ethereum

● Anonymity ● Asymmetric authentication

TABLE I: METHODS OF PREVIOUS WORKS

the citizens who can vote or prevention of the citizens who

already voted for that election. Also, government and citizens

determine the candidates that will be participating in that

election. The ballot box information, candidates and citizen

ballot box relation will be provided by the government which

is the trusted party in the elections. After citizen’s vote, it is

added to the blockchain that we will be proposed below and

any vote has a guarantee from the system about being

immutable. Since a chain contains all the citizen votes

anonymously at the end of the election, the official results will

be announced within minutes after the election terminates.

Any concerning third party can get the chain and count the

votes for being sure that voting is really trusted.

We propose a system that has a leveled structure. There

will be different number of levels in that system according to

necessities of the country. In order to provide a fast,

consistent and secure system, system is designed in a leveled

architecture. This number will change from country to

country according to features of the country. Reasons behind

using a leveled architecture are explained below in detail.

Furthermore, consensus of the system is satisfied using DPoS

algorithms [12].

If the whole country would have been represented with a

single blockchain, synchronization of the system would have

a performance issue due to abundance number of ballots and

the distance between voting centers. Distance in connected

systems is always cause to latency. For a system that includes

all the country under the same blockchain, latency between

two voting center would be a big problem, because for

Turkey, expected latency would be around 100 ms at least.

This is a huge value for a system that consist of ten thousand

of centers and there would be voting at each center

simultaneously. In this case, synchronization of the system

would take lots of time. So, in order to decrease latency,

chains are distributed over levels. From lowest level to

highest level, there will be different chains at each level, and

connections between levels will be provided with a secure

system.

At the lowest level, there will be a chain that consist of

nodes (machines / voting centers) where citizens will perform

their voting about election. Due to the relatively less number

of nodes in the system, synchronization will take affordable

amount of time at the lowest level. When the number of nodes

is arranged in a good pattern (i.e. there will not be overload at

the chains that will cause enormous latency), system will

perform well. Citizen will go to the center and will enter the

system with the identity that is provided by the government.

We considered to build a system that is working on

Figure 1: Use case of the blockchain e-voting system

Figure 2: General system architecture

government’s system that will hold the data about citizen for

specified voting. If a citizen has not voted yet, citizen will be

able to vote one of the candidates. Candidates’ will be hold in

a database that will also be stored at government related

system, because they are already hold. When the

authentication process is satisfied, citizen will vote with

choosing one of the proposed candidates or blank vote for

those who do not want to vote one of the candidates. In this

system, proposed candidates will be taken from database that

includes relation between ballot boxes and candidates. Thus,

there will be only appropriate candidates. In our case, this

government is Government of Turkish Republic. We will be

using e-devlet system [13]. Since most of the authentication

system used across the government related systems are

managed by e-government system, it will not be hard to

implement this system relying on this system. When the user

pass the authentication phase, citizen will see whether he has

voted previously or not. If citizen has not voted yet, citizen

will choose desired candidate according to the steps explained

above.

At the second lowest level, there will be a cluster of chains

that stores data that are coming from below level. In this level,

facilities of blockchain technology are used to make system

consistent. For Turkey, we considered that 2 levels will be

enough. The system at the second level can have about 700

nodes considering population of the country. That brings a

huge performance improvement to the system because the

number of connected nodes decrease in this structure in a

considerable amount. Additionally, if the node numbers at the

2nd or upper levels are increased, performance increases

exponentially. For a country, which has more citizens, level

number can be increased in order to decrease collisions

between transactions. Consequently, system can be

considered as a scalable system.

Communication between levels are ensured using

communication protocols. This communication is need to be

done periodically. So, there will be a time delay between

synchronization of levels. Because, if each vote was

considered instantly, there would be a huge bottleneck. This

synchronization will provide consistency through the system.

For Turkey, according to our calculations, this

synchronization time should be 5 minutes. That means, at the

end of each 5 minutes period, each node cluster will send the

chain data to the upper level node. At this level, data will be

synchronized between nodes using a different

synchronization algorithm. For this level, we designed an

algorithm explained below. You can see the visualization for

this two-leveled example in Fig. 2. As you can see there are

voting centers which are using same blockchain in their

selected area. Also, you can think the voting centers as

numerous voting machines but for the sake of simplicity we

represent them as voting centers. Moreover, you can see that

level 1 nodes are using the same blockchain among the level

1 nodes.

It is stated that vote centers are nodes of blockchains.

There will be a file at each node (voting center) that stores the

number of data that indicates the number of votes accepted

from upper level at previous synchronization step. At every

specified time intervals, voting will be stopped for a very

short time period (it is expected to be 1 minute for Turkey) in

order to synchronize blockchain data between levels. When

the data is arrived to upper level from lower chain nodes, it

will be checked in order to satisfy consistency. If the

consistency of the data is ensured, answer will be a flag that

indicates the data is accepted. At this point, nodes at the level

0 will be waiting an answer from level 1 (and level 1 from

level 2, so on). If arrived flag tells the votes are accepted, the

files at each node (voting centers) will be updated. So, nodes

at the lowest level of one of two levels will always know that

how many votes have been accepted at above level.

Additionally, data will be added to blockchain at the level that

the data is arrived (if the communication is between level 0

level 1, indicated level here is level 1). This data will be

considered as a transaction block, that means all the new votes

(votes coming after from the previously added votes) are

considered as a vote cluster and considered as array in

computer scientific terms. This vote cluster will be a block

that will be added to chain.

At the synchronization phase, if the data coming to the

upper level from different machines are inconsistent, that will

be a case that should be considered with a care. In this case,

if the consensus could not be satisfied, the data will not be

accepted from highest level of the two levels and the

“decline” flag will be sent to the lower one. In this case, same

data should be sent to the upper level again. Until the

consistency is satisfied, this procedure will be continued, and

with this procedure, consistency will be satisfied at each level.

It is stated above that the all nodes at the lower levels

know that the data is accepted if the answer states, so, they

can continue working. But in order to satisfy consistency

through the system, delays between synchronizations should

be arranged very carefully. If the delay between levels

becomes a small amount of time, the time spent for

synchronization may grow much. On the contrary, if the delay

becomes a big amount of time, the data that will be sent

between levels takes an enormous size and to transfer this

data becomes a problem. So, in order to not bottleneck the

system, this delay between levels should be chosen carefully.

With the well-designed synchronization times between

levels, a high performance providing consistent system would

be obtained.

Block structures in levels can be seen below. The election

related data are stored in block as shown below through the

system. As seen in Fig 3. and Fig 4. we propose two types of

blocks: one is for building blockchain at the lowest level of

the architecture, that stores candidate info, ballot box info, a

“nonce” field that will be used for hashing and a prev_hash

info that will be used when creating block that will be added

Figure 3: Block structure at level 1 or more

Figure 4: Block structure at level 0 (lowest level)

to the blockchain, and other one is consisting of two fields:

one is prev_hash field that will be used in order to build

blockchain and other one, “lotb” keyword that indicates list

of the blocks. Attribute “nonce” is a state of art in the

blockchain technology which provides additional security to

each hash. For each chain, you select a pattern such as “hash

must start with 4 trailing zeros”, while creating hash from the

block system look to the hash pattern if it does not fit the

pattern then it changes the “nonce” string until hash pattern is

valid for the chain. It requires more computational power

because it will run hash algorithm multiple times until pattern

is matched but it also increase the security since if any

malicious party create a hash and try to add that block to the

chain, it also has to know the hash pattern. Since chain pattern

will be different for each election, third parties cannot predict

or create a hash that can be accepted.

At the lowest level (level 0), each block will be consisting

of one transaction and at each block, whole related

information about transaction (in proposed e-voting case, this

block indicates a vote) are stored. At the upper levels, the

votes coming from one below level are stored in clusters that

is sent through different time segments, and all of them are

stored as a monolith structure. When this data is arrived, a

new block is processed with the prev_hash data and added as

a block to the blockchain at described level. When a new

block is being tried to add to the chain, according to the

Delegated Proof of Stake of Etherium, it is added and data

consistency is satisfied. Adding block to the chain is very

costly operation. Therefore, there are some implementations

and research about this operation. In [14] Ethereum based

Smart Contracts system has been discussed for different

purposes. Researchers think that smart contracts can be

applied in e-voting and this project implements Smart

Contracts. Smart Contracts reduce the cost of the transactions

and it does not rely on a third party to operate. Turing-

completeness feature of the Ethereum allows to create

customized and more powerful contracts. More technical

details with different inspection methods can be seen in [14].

Summary of the voting procedure can be seen below.

● Authentication: o Node gets credentials ← (voter_id, password)

and sends them with its node_id to e-

government system.

o E-government system validates user credentials, validate (credentials, node_id,

usersInfo).

● Voting: o Let vote be v ← vote (voter_id, candidate) o v is added to blockchain, add (v, chain) o chain information updated for all the voting

machines

o voter’s related field is changed to voted in e- government system, vote (voter_id, userList,

true).

● Counting: o Candidates are received from government,

candidates ← getCandidates(candidateList).

o Using highest level chain, votes are counted and winner of the regions are determined,

results ← count (chain, candidates).

o Final blockchain can be distributed to any third party to inspect the anonym votes with.

IV. ANALYSIS

With the using proposed e-voting system, there will be a

system that can be modified for each election easily. Because

of system is designed considering candidates at the main

voting mechanism and data can be provided however it is

wanted. Thus, this system is easily generalizable.

In this system, whole voting information are hold at the

highest leveled blockchain so, voting information of the

whole country can be reachable instantly at any time after it

is synchronized. So, it will not be a problem to explain results

of the election, and lots of time that is being spent in order to

count votes will be saved. This will change the old and

ineffective system and bring a modern and effective system

to Turkey, and also lead to save lots of energy and money.

According to calculated statistics considering the data

taken from official statistics site of Turkey [15], it is known

that there were about 56 million voters at the 2018 general

elections in Turkey. Considering this number, it is assumed

that 700 clusters (clusters include voting centers under them)

would be enough. This is not a random number; it is chosen

according to population of the country and bottleneck

probabilities at the levels. The geography is a big factor here,

due to the latency.

As all the systems, there is a potential threat for the

proposed election system. As Yli-Huumo et al. states, if any

attacker can get 51% of the computation power of the whole

network then it can manipulate the data and this is called 51%

attack [16]. In proposed system, node can only be gathered

via hacking or stealing the voting machines. Yet, stealing

51% of the voting machines of the country is nearly

impossible because in the elections law enforcements are

protecting the voting centers therefore any physical theft can

be prevented.

Apart from the previously mentioned risks, in case of a

disaster citizen who had to vote on the voting machine that

has damaged or unavailable to server can be distributed to the

nearest voting centers with legal inspections of the supreme

committee of elections. Citizens who vote on the voting

center that has any problem due to technical problems and

disasters can cast their votes in different voting center.

Although disaster can damage to the voting machine,

blockchain keeps secure all the casted votes. Therefore even

in extreme cases during the election day, elections can be

completed safely without any doubt.

V. CONCLUSION AND FUTURE WORK

Democracies depend on trusted elections and citizens

should trust the election system for a strong democracy.

However traditional paper based elections do not provide

trustworthiness. In this paper, we proposed a blockchain

based e-voting system which provides trusted, secure and fast

voting system for Turkey. Proposed system is suitable to

apply in another country whereas integration is a hard work

since each country has different laws and election system

changes between countries. For the future work, system can

be applied for a use case and measurements can be taken to

compare if the calculations hold. Synchronization and

consensus algorithms can be discussed and improved for

better performance and security.

REFERENCES

[1] R. Hanifatunnisa and B. Rahardjo, "Blockchain based e-voting recording system design," 2017 11th International Conference on Telecommunication Systems Services and Applications (TSSA),

Lombok, 2017, pp. 1-6.

[2] R. Krimmer, A. Ehringfeld, and M. Traxl, “The Use of E-Voting in the Austrian Federation of Students Elections 2009,” Internet:

https://pdfs.semanticscholar.org/6b8f/34a5bd3e7eabc7e3a9a3f008187

e4415e26a.pdf [Nov. 26, 2018]

[3] “The Geneva Internet Voting System,” Internet: https://www.coe.int/t/dgap/goodgovernance/Activities/E- voting/EVoting_Documentation/passport_evoting2010.pdf [Nov. 25,

2018]

[4] F. Þ. Hjálmarsson, G. K. Hreiðarsson, M. Hamdaqa and G. Hjálmtýsson, "Blockchain-Based E-Voting System," 2018 IEEE 11th

International Conference on Cloud Computing (CLOUD), San

Francisco, CA, 2018, pp. 983-986.

[5] S. Ølnes, J. Ubacht and M. Janssen, "Blockchain in government: Benefits and implications of distributed ledger technology for

information sharing", Government Information Quarterly, vol. 34, no. 3, pp. 355-364, 2017.

[6] A. Barnes, C. Brake, and T. Perry, “Digital Voting with the use of Blockchain Technology,” Available:

https://www.economist.com/sites/default/files/plymouth.pdf [Nov. 20,

2018]

[7] A. Barnes, C. Brake, and T. Perry, “Digital Voting with the use of Blockchain Technology,” Available: https://www.economist.com/sites/default/files/plymouth.pdf [Nov. 20,

2018]

[8] M. Pawlak, A. Poniszewska-Marańda and N. Kryvinska, "Towards the intelligent agents for blockchain e-voting system," Procedia Computer

Science, vol. 141, pp. 239-246, 2018.

[9] P. Tarasov and H. Tewari, "The Future of E-voting," IADIS International Journal on Computer Science and Information Systems,

vol. 12, no. 2, pp. 148-165.

[10] Bartolucci, S., Bernat, P., & Joseph, D. (2018). SHARVOT: Secret SHARe-Based VOTing on the Blockchain. 2018 IEEE/ACM 1st

International Workshop on Emerging Trends in Software Engineering for Blockchain (WETSEB), 30-34.

[11] A. Reyna, C. Martín, J. Chen, E. Soler and M. Díaz, "On blockchain and its integration with IoT. Challenges and opportunities", Future

Generation Computer Systems, vol. 88, pp. 173-190, 2018.

[12] B. Wang, J. Sun, Y. He, D. Pang and N. Lu, "Large-scale Election Based On Blockchain", Procedia Computer Science, vol. 129, pp. 234-

237, 2018.

[13] Internet: https://lisk.io/academy/blockchain-basics/how-does- blockchain-work/delegated-proof-of-stake [Nov. 25, 2018]

[14] Internet: https://www.turkiye.gov.tr/ [Nov. 25, 2018]

[15] Alharby, Maher, and Aad van Moorsel. “Blockchain Based Smart Contracts : A Systematic Mapping Study.” Computer Science & Information Technology (CS & IT), 2017

[16] Internet: https://sonuc.ysk.gov.tr/ [Nov. 16, 2018]

[17] J. Yli-Huumo, D. Ko, S. Choi, S. Park and K. Smolander, "Where Is Current Research on Blockchain Technology?—A Systematic

Review," PLOS ONE, vol. 11, no. 10, 2016.

View publication statsView publication stats