INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE

O. Oksuz, Vol.11, No.1, pp.1-14

Preserving Identity Leakage, Data Integrity and

Data Privacy Using Blockchain in Education
System

Ozgur Oksuz 

Adiyaman University
Department of Computer Engineering
Adiyaman, TURKEY
[email protected]
Research Paper Received: 30.01.2022 Revised: 26.03.2022 Accepted: 03.04.2022

Abstract—Today, blockchain technology is evolving and has been used in many sectors such as healthcare, supply chain
management, internet of things (IoT) and cryptocurrency exchange. Using this technology in these areas provides very good
functionalities. A blockchain network is immutable, public, open, distributed, secure and reliable. This paper is about using
blockchain technology in education system. Applying blockchain technology to education system brings all those properties above.
However, user (student) data privacy and identity management of the entities in the system should be also considered. In this paper,
the proposed scheme not only satisfies all these properties but also protects student data privacy and identity management of the
entities when they communicate with each other. The proposed construction consists of encryption algorithms to protect students’
private data and provide secure communication between the entities. Moreover, the proposed scheme does not leak students’
identities to third parties in the blockchain.

Keywords—blockchain, data privacy, identity leakage, data integrity.

1. Introduction 51% honest nodes exists), and reliable (the network

is up all the time). It is used as a building block
Blockchain network was firstly introduced and in many systems such as healthcare [2], internet of
used as a building block for Bitcoin [1]. Then, the things [3], and supply chain management [4].
importance of it has been evolving since it has very
useful properties. A blockchain network is public In this paper, the blockchain technology is used
(all transactions are publicly available), open (any- in education system. With this technology many
one can participate in the network), decentralized (it disadvantages in education system are eliminated.
does not require a trusted party for authorization), One of the disadvantages is to generate fake degree
distributed (each transaction’s validity is performed certificates [5]. Moreover, providing false informa-
many other nodes in the network), secure (if at least tion about academic achievements/transcripts is also

1
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

another problem when a student applies for an the blockchain. Student’s data integrity is pre-
internship/summer job. Furthermore, storing each served.
student’s academic information/records for a long The organisation of the paper as follows: Section
time can create another problem. These records 2 presents related work. Section 3 introduces the
can be lost due to human mistake or to have definitions that are going to be used throughout
some natural disasters such as flooding. Putting all the paper. In Section 4, the architecture and the
students’ academic records into blockchain elimi- transaction types are introduced in the blockchain.
nates these kinds of disadvantages. However, other Then, the proposed scheme is presented in Section
problems can arise. One of the problems is to 5. In Section 6, security analysis of the scheme is
protect student’s data privacy. Data privacy says that given. In Section 7, we give the conclusion.
any unauthorized entity cannot read the student’s
sensitive information such as student’s real identity, 2. Related Work
course grades, taken courses, enrolled program, and
owned certificates. Thus, data privacy should also There have been some studies that use blockchain
be addressed using blockchain. Another issue that technology in education system. These studies are
should also be considered is identity leakage. The [6], [7], [8], [9], [10], [11]. In [6] students’ grades
students’ identities should be hidden in the sys- are not in the blockchain. The completed courses
tem. Otherwise, any untrusted third party can map of the student are putting into the blockchain.
transactions in blockchain to a student. Then, this However, the courses are still disclosed. This may
untrusted party can do some statistical analysis to rise a problem that student can be rule out if it
make a profile for the student. To address the above is overqualified. Our scheme does not leak any
problems, this paper introduces a system that has sensitive information about student at the beginning.
the following properties: Our scheme leaks only the required courses based
on the job title. The study in [8] uses two kinds
• A brief mathematical formalization and con- of blockchains: private and public. Moreover, it
struction of the system is given to protect sensi- focuses on privacy and integrity of the students’
tive data of the students, manage identity of the data. [9], [10] focus on putting students’ certificates
parties, and process of adding student records into the blockchain not the courses and grades of the
to the blockchain. students. In [11], the first university (the university
• The sensitive data (student’s real identity, grade, of Nicosia) issues the academic certificates for stu-
course identity, program identity) of the stu- dents. The given certificate is verified using bitcoin
dents is stored in the blockchain and is en- blockchain. In [7], Sony Global Education develops
crypted. The data is only seen in the clear a new blockchain for storing academic records. The
by the third parties when a student is trans- work in [12] presented a blockchain-based (permis-
ferred/enrolled to another university/program or sioned) repository for educational credentials but it
applies to an internship. In this case, the leakage does not have privacy protection of the students’
is only limited to the corresponding transactions private data. The work in [13] has the same problem
based on the requirements of the position that that it does not introduce any protection mechanism
the student applies to. for students’ private data. The studies in [14], [15],
• The data of all transactions are immutable in [16], [17], [18] use permissionless blockchain in

2
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

education system. [15] does not have protection an output y is given, it is infeasible to compute x
mechanism against student private data. Studies from y. This property is called as ”One Way”. It
in [15], [16], [17], [18] use off-chain storage. It is infeasible to find another z that satisfies H(x) =
means that students’ credentials are not stored in H(z) = y. This is called as ”Collision Resistant”.
blockchain. They are stored in a file storage system. A secure hash function algorithm should satisfy the
Indy Hyperledger [19] is a public permissioned all above properties. SHA-256 can be used to have
blockchain system that allows users to share all these properties.
their identities (based on credentials) anonymously.
Moreover, this system is able to issue and revoke 3.2. Symmetric/Private Key Encryption
cryptographic credentials [20], [21]. Anyone with
read access to the ledger can verify signatures made A symmetric encryption key scheme consists of
by issuers on credentials. In order to use this system 3 algorithms: Key generation (KGen), Encryption
(hyperledger Indy), the credentials that students own (SEnc), and Decryption (SDec). KGen is the
need to be known in advance. Every user can have key generation algorithm takes a security param-
different credentials. However, it could be very eter and outputs a key for user u: Secret key
difficult to manage these credentials. In addition, (SSku ). Encryption algorithm takes a message m
it uses heavy cryptographic operations. Moreover, and the secret key (SSku ), outputs the ciphertext
in this work, these kind of heavy cryptographic C = SEncSSku (m). Decryption algorithm takes
tools are not used. The students do not participate the secret key SSku and C, outputs message m =
in the blockchain network and the blockchain does SDecSSku (SEncSSku (m)). A secure symmetric key
not contain the students’ identities in the clear. encryption scheme should be resilient at least to
This hides students’ identities in the blockchain. Chosen Plaintext Attack (CPA). It means that even
However, once the student applies/communicates to a message m is encrypted over and over again,
a company for an internship or a permanent job, the the resulting ciphertext should be different each
student’s real identity is disclosed to corresponding time. An example of CPA secure algorithm is AES
recruiter. The recruiter needs to know the student to (Advanced Encryption Standard) with CBC mode.
communicate with the institution for this student to
verify if the student has satisfied the internship/job 3.3. Asymmetric/Public Key Encryption
requirements.
An asymmetric encryption protocol consists of
3. Definitions 3 algorithms: Key generation (KGen), Encryption
(P Enc), and Decryption (P Dec). KGen is the key
3.1. Hash Function generation algorithm takes a security parameter and
outputs a key pair for user u: Public key P P ku
A hash function is a mathematical algorithm that and Secret key P Sku . Encryption algorithm takes
has the following properties: It maps an input x of a message m and public key P P ku , outputs a
any size to an output of fixed size y. This is shown ciphertext C = P EncP P ku (m). Decryption algo-
as H(x) = y. It is deterministic algorithm, which rithm takes the secret key P Sku and C, outputs
means that if the algorithm retakes x, the output is message m = P DecP Sku (P EncP P ku (m)). A secure
always y. The algorithm is efficiently computed. If asymmetric key encryption scheme also should be

3
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

resilient at least to Chosen Plaintext Attack (CPA).

Examples of CPA secure schemes are RSA (not
the textbook) [22], and Paillier encryption schemes
[23].

3.4. Signature

A digital signature protocol consists of 3 al-

gorithms: Key generation (KGen), SIGN , and Figure 1. Network architecture of the proposed
V ERIF Y . KGen is the key generation algorithm scheme.
takes a security parameter, outputs a key pair
for user u: Signature public key/verification key
If the transaction is intact, then it is added to the
(SignP ku ) and Signature secret key (SignSku ).
ledger. In the ledger, transactions are kept as blocks.
SIGN algorithm takes a message m and signature
It uses mathematical structures (hash functions and
secret key (SignSku ), outputs a signature S =
signatures) to verify each transaction. The first block
SIGNSignSku (m). V ERIF Y algorithm takes sig-
is always named as genesis block. This is a special
nature public key SignP ku , signature S, and mes-
block.
sage m, outputs 1 (V ERIF Y (m, S, SignP ku ) ==
1) if the signature is generated by user u under
message m. If V ERIF Y (m, S, SignP ku ) == 0, 4. Architecture of the Proposed Scheme
then the signature under message m is not generated
by user u. The proposed scheme consists of 4 different ac-
A secure signature scheme needs to satisfy two tors: a Central Authority (CA), Institutions (Ins),
properties: authenticity and integrity. Authenticity Students (Stu) and Recruiters (Rec). CA is re-
says that the owner of the signature convinces a sponsible for setting up the system P KI (public
verifier that the owner of the signature generates key infrastructure). Moreover, CA also authorizes
the signature using the message. Integrity says that undergraduate courses and degrees. CA can be gov-
signed data cannot be altered by any entity. An ex- ernment that everyone can rely on. The institutions
ample of secure signature algorithm can be Schnorr are the places that students take courses and get
signatures [24]. degrees. Institutions can be schools, universities.
Students take courses from schools/institutions. Re-
3.5. Blockchain cruiters are the employers that wants to hire quali-
fied students. Moreover, the recruiters should also be
Blockchain is a technology that it has the fol- convienced that the students satisfy the requirements
lowing properties: It is a peer-to-peer decentralized of the positions.
system that nobody controls the system. It has The network architecture of the proposed scheme
the ledger technology that it keeps the information is shown in Fig.1. In the system, there are two kinds
(transaction) up to date and it is immutable. Each of nodes: trusted and untrusted. The trusted nodes
node (computer) keeps a copy of this ledger. Since it can only issue transactions into a ledger. After a
is decentralized, other nodes check each transaction. block is formed, the block is published and being

4
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

public.
The work flow of the system is the following:
0. Each node (trusted or untrusted) registers its
public key to CA. Then, CA authorizes each
node’s public key. Moreover, CA makes these
public keys public. Thus, everyone in the sys-
tem knows the identities of others.
1. After grading a student’s exam, an institution Figure 2. An illustration of storage, course
sends the student’s grade as a transaction to policy and program policy blocks/transactions.
other trusted nodes.
2. If most of the nodes (more than 50%) verify
that the transaction is correctly formed, it is put an internship. Students can present fake information
into the blockchain as a block by every trusted about their academic history when they apply a job
node. or an internship.
3. If a Stu is transferred to another institution or We assume that malicious adversaries can not
applying on a job, an Ins or a Rec asks the collude with each other. It means that they can
current institution to hand over some tokens not share specific information with each other. This
to retrieve the student information from the information can be their real identities, public keys
blockchain. (pseudonym identities) and their secret keys. More-
4. Once an Ins or a Rec gets corresponding over, any student can not collude with any recruiter.
tokens from the student’s current institution,
the Ins or the Rec retrieves the corresponding
4.2. Transactions
student’s information (as a transaction) from the
blockchain.
In this section, transactions (storage, course policy
and program policy) of the proposed scheme as
4.1. Threat Model
blocks are going to be presented. In other words,
each transaction is going to be a block. In Fig.2,
In this paper, CA and Institutions are trusted en- there are three types of transactions. Multiple trans-
tities. Recruiters and students are untrusted entities. action types in a blockchain are not new. These
Recruiters and students are malicious adversaries are used in other studies such as in [2]. While
that they can provide fake or altered information institutions issue storage transactions, course policy
to institutions. Recruiters try to learn private data transactions and program policy transactions are
of students in the blockchain. Recruiters can ask only issued by CA.
institutions for decryption keys to learn specific
A storage block consists of the following infor-
students’ private data in the blockchain. Moreover,
mation:
recruiters can change data that comes from student
A. Then, it asks institution for decryption keys to S1. Transaction Type (T T ): This is type 00 which
learn private data of student B. Student A can also is storage transaction.
pretend to be student B when applying a job or S2. Transaction Number (T N )

5
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

S3. Student’s Real Identity (SRI): This identity CA.

consists of student’s national identity number, CP8. Time Stamp (T S): Time that the transaction
full name, date of birth, and gender. occurs. This is done by CA.
S4. Course ID Number (CID): This number is CP9. Data Hash (DH): Hash of all the data (step
going to be course number that a student takes. CP1-CP8). This is done by the CA.
S5. Program ID Number (P ID): The number of CP10. CA Signature for Data Hash (CASIGN ):
the program that the student enrols. This is for data integrity. CA signs the data in
S6. Grade (G): This is a number that is going to step CP9.
be between 0 and 100.
In step CP5, this structure is important when a
S7. Institution ID Number (IID): This is the sig-
student decides to enrol another program of the
nature public key of the institution.
same university or a different university. The graph
S8. Hash of the Previous Block (HP B): This is
shows if the course is offered by different programs.
for immutability for the blockchain. The hash
Thus, the student does not need to retake the course.
of the previous block is added. This is done by
This provides minimum of difficulty when student
an Ins.
change its major. As an example, introduction to
S9. Time Stamp (T S): Time that the transaction
computer programming course is offered to many
occurs. This is done by an Ins.
other programs such as Mathematics, Computer
S10. Data Hash (DH): Hash of all the data (step
Science, Civil Engineering and so on. If a student
S1-S9). This is done by an Ins.
decides to change its major from Mathematics to
S11. Institution Signature for Data Hash
Computer Science, the students only need to transfer
(InsSIGN ): This is for data integrity.
its grade. It does not need to retake the same
The institution signs the data in step S10.
course (introduction to computer programming). An
A course policy transaction/block consists of the illustration of a graph structure is shown in Fig.3. In
following information: the figure, Mathematics program is represented by
146729, Computer Science program is represented
CP1. Transaction Type (T T ): This is type 01 which by 190834, Environmental Engineering program is
is a course policy transaction. represented by 146892, Electric Engineering pro-
CP2. Transaction Number (T N ) gram is represented by 154266, and Introduction to
CP3. Course ID Number (CID): This number is the Programming course is represented by 034783.
going to be course number that a student takes.
A program policy transaction/block consists of the
CP4. Course Requirements (CR): It gives course
following information:
information such as number of credits, course
syllabus and minimum score to pass the course. PP1. Transaction Type (T T ): This is type 10 which
CP5. Graph Structure (GS): It shows which other is a program policy transaction.
programs offer this course. PP2. Transaction Number (T N )
CP6. CA ID Number (CAID): This is the signa- PP3. Program ID Number (P ID): This number is
ture public key of CA. going to be program number that a student
CP7. Hash of the Previous Block (HP B): This is enrols.
for immutability for the blockchain. The hash PP4. Program Requirements (P R): It includes total
of the previous block is added. This is done by credits that student needs to take to complete

6
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

need to be authorized to join the blockchain net-

work. Permissionless blockchain network is used in
Bitcoin [1] and Etherium [25]. In a permissioned
blockchain, not anyone can join the network. Per-
mission is provided to certain identifiable partici-
pants to join the network. This requirement adds an
additional level of security. Participants are known
to each other. The trusted parties only write to
Figure 3. An illustration of a course-program the ledger. Permissioned blockchain is also used
relation graph. in many protocols since permissioned blockchain
has an additional level of security. Moreover, per-
missioned blockchain is more efficient and more
the program, program information, program scalable. Since in our protocol, the entities are
courses and their numbers. known in advance we use permissioned blockchain
PP5. CA ID Number (CAID): This is the signa- network. CA and Ins are allowed to write data to
ture public key of CA. the ledger. The scheme enjoys with the following
PP6. Hash of the Previous Block (HP B): This is properties:
for immutability for the blockchain. The hash
• The sensitive data (student’s real identity, grade,
of the previous block is added. This is done
by CA. course identity, program identity) of the stu-
PP7. Time Stamp (T S): Time that the transaction dents are encrypted in the blockchain. More-
occurs. This is done by CA. over, there is no information about identitites of
PP8. Data Hash (DH): Hash of all the data (step the students in the clear. This information is hid-
PP1-PP7). This is done by the CA. den. The data is only seen by the third parties in
PP9. CA Signature for Data Hash (CASIGN ): the clear when a student is transferred/enrolled
This is for data integrity. CA signs the data to another university/program or or applying to
in step PP8. an internship. The only leakage is based on the
required job position when the student applies
for or courses that need to be transferred when
5. Construction the student is transferred to another institution.
• The data integrity is preserved and cannot be
The proposed protocol uses blockchain network. changed. The data of all transactions are im-
There are two kinds of blockchain network that are mutable in the blockchain ledger.
open to public: Permissionless and Permissioned. In
a permissionless blockchain, anyone can join and The proposed protocol consists of 4 algorithms:
leave the network. When a participant joins the Setup, Register, T ransaction, and Key T ransf er
network, it can have the copy of entire blockchain f or Decryption.
and operate as a full node. Anyone can validate Setup: In this phase, CA decides which signature
the transactions publicly. All data in the blockchain (SIGN ), public/symmetric key encryption algo-
is available to everyone. There is no trust between rithms (P Enc/SEnc) and their appropriate security
the participants. Moreover, a participant does not parameters will be used in the system. Once the

7
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

algorithms are announced to the public, each node with each other, they show their valid certificates –
(trusted and untrusted) and CA chooses their secret that are digitally signed by CA’s private key- to each
key/public key pair for signature algorithm and other.
chose their secret/public key pair for public key
Moreover, each student Stuq registers its signature
encryption algorithm. CA’s secret key and public
public key SignP kstuq and public encryption key
key pair for signature algorithm are represented as
P P kStuq to its corresponding institution Insj . This
SignSkCA and SignP kCA . A trusted node i’s secret
can be done when student enrols a program in that
key and public key pair for signature algorithm
institution.
are represented as SignSki,t and SignP ki,t . An
untrusted node j’s secret key and public key pair for Transaction: Three cryptographic primitives are
signature algorithm are represented as SignSkj,ut used for storage transaction phase: Symmetric En-
and SignP kj,ut . cryption (SEnc), Hash (H) and Signature (SIGN ).
Once student’s exam is graded, institution j (Insj )
Moreover, CA’s secret key and public key pair
encrypts its grade G, corresponding real identity
for public key encryption algorithm are represented
(SRI) of the student, corresponding course identity
as P SkCA and P P kCA . A trusted node i’s secret
number CID and corresponding program identity
key and public key pair for public key encryption
number P ID. We use the following notations: GStuq
algorithm are represented as P Ski,t and P P ki,t . An
as the grade of student q, SEnck (GStuq ) as the en-
untrusted node j’s secret key and public key pair
cryption of the grade G of student q, SEnck′ (CID)
for public key algorithm are represented as P Skj,ut
as the encryption of the course identity number
and P P kj,ut .
CID, SEnck′′ (P ID) as the encryption of the pro-
Furthermore, each student Stuq also chooses se- gram identity number P ID and SEnck′′′ (SRI)
cret key and public key pair for signature algo- as the encryption of the student’s real identity.
rithm as SignSkStuq , SignP kStuq and chooses pub- Here k, k ′ , k ′′ , k ′′′ are four different secret private
lic/secret key as P SkStuq , P P kStuq for asymmetric encryption keys chosen randomly by Insj for SEnc
public key encryption algorithm. A note that all algorithm. Moreover, Insj uses hash algorithm H
secret keys are chosen randomly for security. to hash the following data
Register: In this phase, each node’s (trusted or un- DH = H(T T ||T N ||SEnck′′′ (SRI)||SEnck′ (CID)||
trusted) public encryption key P P ki and signature
SEnck′′ (P ID)||SEnck (Gstuq )||SignP kj ||HP Bl ||T S),
public key SignP ki for node i are going to be
where ∥ is the concatenation symbol.
registered to the system by CA. In the system, each
public key (for signature and encryption) will be In the hash data, institution signature public key
identity of a party (trusted node, untrusted node, SignP kj is chosen as the institution identity num-
CA). A standard Public Key Infrastructure (PKI) ber. Moreover, HP Bl is chosen as hash of the
scheme can be used to register entities public keys. previous block. We assume that current block is
CA also acts as certificate authority that is respon- l + 1. We also assume that SRI is only known to
sible for identities of the parties in the blockchain student itself and the institution. This information
network. After the parties’ identities have been is never seen in public. Only authorized people
verified, CA generates digital certificate under the (old and new institution, recruiter when student
parties’ public keys. When the entities communicate applies a job or an internship, and CA) can know

8
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

this information. Finally, Insj signs DH value P R, P P 5 = CAID, P P 6 = HP Bf , P P 7 = T S,

as SIGNSignSkj (DH) and generate the transaction P P 8 = DH, P P 9 = SIGNSignSkCA (DH).
that consists of the followings:
In the transaction, T T is set to 10 since it is
S1 = T T , S2 = T N , S3 = SEnck′′′ (SRI), a program policy transaction. We assume that the
S4 = SEnck′ (CID), S5 = SEnck′′ (P ID), S6 = current (this case) block is f + 1. This case DH is
SEnck (Gstuq ), S7 = SignP kj , S8 = HP Bl ,
DH = H(T T ||T N ||P ID||P R||CAID||HP Bf ||T S).
S9 = T S, S10 = DH, S11 = SIGNSignSkj (DH).
T T is set to 00 since it is a storage transaction. As a note that course policy transactions and
Once the transaction is generated by Insj , Insj program policy transactions are done by CA so
sends this transaction to the other trusted nodes. these transactions can be done before any storage
Other trusted nodes check the signature and data transactions.
if they pass from verification of the signature. If Key Transfer for Decryption: In this phase hap-
more than 50% of the trusted nodes respond that pens when a student is transferred to another insti-
the signature is valid, this transaction is added to tution or student applies a job or an internship. An
the ledger. illustration of the message flow diagram is given
As a note that noone can see the student’s real in Fig. 4. The student sends a message tuple to its
identity, grade of the student, the student’s course current institution when it applies to an internship or
information and program information in the clear in a job (S3). Assuming that Stuq ’s current institution
the transactions since these are all encrypted. Thus is Insj and applying for an internship in Recm .
any unauthorized party cannot see the plaintexts of The student sends the same message tuple to the
them. Since SRI is encrypted so any unauthorized recruiter by mentioning its current institution and
entity cannot map this encrypted identity to a real its identities (S4).
person. Furthermore, there is no information about Since the blockchain is public, recruiter Recm can
student in the clear in the blockchain. So the iden- see all the transactions but cannot see the private
tities of the students are hidden. student information. This is because all sensitive
A course policy transaction phase consists of information about the student is encrypted. In order
CP 1 = T T , CP 2 = T N , CP 3 = CID, to see sensitive student information, Recm needs
CP 4 = CR, CP 5 = GS, CP 6 = CAID, to know the encryption keys to decrypt the corre-
CP 7 = HP Bt , CP 8 = T S, CP 9 = DH, sponding ciphertexts. Moreover, the corresponding
CP 10 = SIGNSignSkCA (DH). transaction numbers needs to be known by Recm
In the transaction, T T is set to 01 since it is since the blockchain data does not contain any
a course policy transaction. We assume that the student identity in the clear. Thus Recm asks the
current (this case) block is t + 1. This case DH student’s institution Insj to send the corresponding
is transaction keys and the corresponding transaction
numbers (R2). These keys are needed for decrypt-
DH = H(T T ||T N ||CID||CR||GS||CAID||HP Bt
ing the student’s sensitive information (student’s
||T S). real identity, grades, course identities and program
A program policy transaction phase consists of identity) that they are required for the position
P P 1 = T T , P P 2 = T N , P P 3 = P ID, P P 4 = that student applies. For example, if the student

9
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

the signature of Reqs that it is signed by Recm .

RecruiterInf o can be the new institution
information that student is transferred to. We
assume that student applies to an internship
position. The student prepares the tuple M =
(Internship||SignP kInsj ||SignP kRecm ||
SignP kStuq ||SRI||Reqs).
S2. Stuq computes P EncP P kInsj (M ) and
Figure 4. Message Flow Diagram between Par- signs it with its signature secret key
ties. SIGNSignSkStuq (P EncP P kInsj (M )),
S3. Stuq sends P EncP P kInsj (M ) and
SIGNSignSkStuq (P EncP P kInsj (M )) to Insj .
wants to be an intern in a software company. The
This is for letting Insj know that Stuq wants
requirements of the internship for the student are
to apply an internship. Moreover, the recruiter
to take some programming language courses. Thus,
is Recm .
the student needs to prove that it has taken some
S4. Stuq also computes P EncP P kRecm (M )
programming language courses. Some projects can
and SIGNSignSkStuq (P EncP P kRecm (M )).
also be given related to software used by the student
Then Stuq sends them to Recm . With this
during the projects. This information is all encrypted
information, Stuq applies to the internship
so Insj needs to send all decryption keys with the
position.
transaction numbers to Recm (I3).
Student Recruiter

S1. Stuq generates a tuple, R1. Recm first decrypts P EncP P kRecm (M ) in step-
(App||InstitutionInf o||RecruiterInf o|| S4 as DecP SkRecm (P EncP P kRecm (M )) and gets
StudentInf o1||StudentInf o2||Reqs), where the tuple (M ). Thus, Recm knows Stuq ’s real
App ∈ {Job, Internship, T ransf er} identity SRI, pseudonym identity SignP kStuq
shows the purpose of the application. and student’s current institution information.
InstitutionInf o is the current institution Recm also checks if the signature is a valid
information which is SignP kInsj (signature signature.
verification key of the institution), R2. Recm knows who is applying to the po-
RecruiterInf o is the recruiter information sition. If the signature is valid from step-
which is SignP kRecm (signature verification R1, Recm computes P EncP P kInsj (M ) and
key of the recruiter), StudentInf o1 is SIGNSignSkRecm (P EncP P kInsj (M )). Then it
the student’s pseudonym identity which is sends them to Insj .
SID = SignP kStuq (signature verification key
Institution
of the student), StudentInf o2 is the student’s
real identity which is SRI, and Reqs is I1. Insj firstly decrypts the ciphertext in step-S3
the requirements of the position of a job/an and gets M . Then, it checks/verifies the sig-
internship. Reqs consists of the requirements nature using SignP kStuq that if the encrypted
of the position, deadline of the application and message (in step-S3) is generated by Stuq .

10
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

If the signature is valid, Insj learns what course. In the block, the recruiter simply retrieves
Stuq wants to do (internship/job application or CP4 (course requirements). This allows the recruiter
transfer to another university) and where Stuq to see what topics have been covered in this course.
wants to apply (who is the recruiter) and what
the requirements are for the position that Stuq 5.1. Discussions
applies (Reqs). These are all learned from M .
Insj also checks if SRI and SID belong to When a student wants to be transferred another
same student. institution and wants to study the same program, the
I2. Next, Insj decrypts the ciphertext in step- new institution requires all the course history of the
R2 and gets M . Then, it checks/verifies the student. This case old university needs to send all
signature that if the encrypted message (in step- the secret keys of the corresponding ciphertexts with
R2) is generated by Recm . all the corresponding transaction numbers. When a
I3. If the tuples from step-I1 and step-I2 are the student applies for a job that it requires completing
same and all the checking/verification steps in the program of study, the recruiter (employer) also
I1 and I2 are successfully completed, Insj needs to have all course history of the student with
sends transaction numbers (T N ) and the se- the transaction numbers. This is because whether
cret keys to decrypt ciphertexts (encrypted stu- the student has completed the all the requirements of
dent’s private information) in the transactions the program for a degree. For this case, the recruiter
to Recm . needs to retrieve program policy transaction in the
blockchain for verification.
Once Recm has the secret keys, it can decrypt the
sensitive data required for the position. For example, A student can provide fake academic history via
if Insj sends its CV/Resume to the recruiter when the student
applies for the position. This can be done by the
P EncP P kRecm (IID||T N ||k ′′′ ||k ′ ||k ′′ ||k) and
student updating the message tuple as follows:
SIGNSignSkInsj (P EncP P kRecm (IID||T N ||k ′′′ ||k ′ ||k ′′
(Internship||SignP kInsj ||SignP kRecm ||SignP kStuq
||k)) to Recm . Recm decrypts the ciphertext as
||SRI||CV /Resume||Reqs). Since all the academic
P DecP SkRecm (P EncP P kRecm (IID||T N ||k ′′′ ||k ′ || history of the student is stored in the blockchain
′′ ′′′ ′ ′′
k ||k)) = IID||T N ||k ||k ||k ||k. It gets tuple and it is immutable, student does not get any
(IID||T N ||k ′′′ ||k ′ ||k ′′ ||k). Then, it checks if the advantage by presenting fake results (CV ) to the
signature is valid using verification key IID = recruiter. The recruiter needs to communicate with
SignP kInsj . If the signature is valid, it decrypts the institution for the student’s real academic history
the ciphertexts in the transaction using the keys for the requirements (Reqs) of the position. Thus,
k, k ′ , k ′′ , k ′′′ to retrieve the student’s information. the recruiter can easily verify student’s CV this
After decrypting all the sensitive information, the case. Since the student can not get any advantage
recruiter sees the student’s grades, course identities, by faking its academic history, we don’t add the
program identity, real identity. Moreover, once the student’s CV to the tuple in the real protocol.
recruiter gets course identity, it can look for the As a note that students can study multiple pro-
course identity in the course policy transaction block grams, masters and PhDs. Moreover, students can
in the blockchain for getting information about the enroll multiple programs and have multiple degrees,

11
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

and have their certificates. Putting these certifi- in a company (recruiter), the leakage is limited that
cates as pictures and all other information to the only the required decryption keys of the transac-
blockchain can be problematic. This is because each tions and corresponding secret keys are sent to the
node in the blockchain network has limited source recruiter. Moreover, the decryption keys are not sent
capabilities such as storage. This case the values in the clear, they are sent in the encrypted form.
should be stored in other sources (off-chain) such Thus, only the authorized recruiter can recover the
as in a cloud [2] or in a central database [26]. keys by decryption. Then, the recruiter can recover
For these cases, the encrypted private data of the the student’s real identity, grade, program identity
students are stored in a database, the addresses and course identity by using those keys.
of these data are stored in the blockchain. After Data Integrity and Data Immutability: Data in-
the student applies for an internship, the institution tegrity is observed by using signature algorithm
not only sends the decryption keys but also sends in the proposed scheme. With this algorithm the
the transaction numbers to the recruiter. Once the data cannot be altered by any entity. The data
recruiter gets the transaction numbers, it looks for immutability is preserved in the construction by
the transaction numbers in the blockchain to get using blockchain technology. With this technology
the addresses of the encrypted student’s data. Then, each block (transaction) is tied each other with the
the recruiter retrieves the encrypted student’s data help of signature and hash algorithms. Thus, no one
from the database using the addresses. Using the can break this chain to change data. Moreover, a
decryption keys, the recruiter gets the plaintexts of malicious recruiter can not gain any useful infor-
the data. mation by changing student A’s message to student
B’s message. The malicious recruiter can do this
6. Security Analysis to get decryption keys to learn student B’s private
data. This is because institution needs to have the
User/student Data Privacy: User/student’s sensi- appropriate message from student B too. In other
tive data consists of Student’s grade, Student’s real words, steps S3 and R2 should be consistent. A
identity, Course identity number and Program iden- student can not pretend being another student when
tity number. These data are very sensitive. Leaking it generates a message since the student needs to
this information in the clear results breaching of know another student’s real identity and pseudonym
student’s data privacy. However, any unauthorized identity and signature secret key. The student can
entity cannot see these data in the clear. This in- not provide false information about its academic
formation is encrypted using secure symmetric key history since all the academic history is stored in
encryption scheme such as AES with CBC mode. the blockchain by the trusted parties (institutions).
AES − CBC mode consists of an initialization
vector (IV ) as a random number in the algorithm
that even the same message is encrypted more than 7. Conclusion
once, the resulting ciphertext is going to be different
each time. Thus the proposed scheme provides data In this paper, a protocol is introduced for an
privacy of the student. In the case when the student education system using blockchain technology that
is transferred to another institution to study different provides student data privacy, data integrity, and
program or when the student wants to do internship immutability of the data. The proposed construction

12
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

uses encryption algorithms (symmetric and asym- [7] S. G. Education, “Sony global education develops technology
metric) to protect data privacy, uses signatures for using blockchain for open sharing of academic proficiency and
progress records,” Tech. Rep., 2016.
data integrity and uses blockchain for immutabil- [8] K. Kuvshinov, I. Nikiforov, J. Mostovoy, D. Mukhutdinov,
ity. Moreover, the proposed scheme does not leak K. Andreev, and V. Podtelkin, “Disciplina: Blockchain for
students’ real identities in the blockchain since the education,” Yellow Paper, Tech. Rep., 2018.
[9] R. Arenas and P. Fernandez, “Credenceledger: A permissioned
students do not need to participate in the blockchain
blockchain for verifiable academic credentials,” in 2018 IEEE
network. In addition, all the sensitive data is en- International Conference on Engineering, Technology and In-
crypted. The proposed protocol models commu- novation (ICE/ITMC), Stuttgart, Germany, 17-20 June 2018, pp.
1–6.
nication between the parties and proposes iden-
[10] Blockcerts, the open standard for blockchain certificates.
tity management. This paper also proposes course- http://www.blockcerts.org/. Accessed April 5, 2022. [Online].
program graph for the scheme to provide minimum Available: http://www.blockcerts.org/
of difficulty when the student change its major. [11] University of nicosia. blockchain certificates (academic
and others). https://www.unic.ac.cy/iff/blockchain-certificates.
Accessed April 5, 2022. [Online]. Available: https:
//www.unic.ac.cy/iff/blockchain-certificates
Acknowledgments
[12] E. E. Bessa and J. S. Martins, “A blockchain-based educational
record repository,” 2019, arXiv preprint arXiv:1904.00315.
The author thanks the anonymous reviewers for [13] M. Sharples and J. Domingue, “The blockchain and kudos:
A distributed system for educational record, reputation and
their useful comments and suggestions. reward,” in Adaptive and Adaptable Learning, K. Verbert,
M. Sharples, and T. Klobučar, Eds. Cham: Springer Inter-
national Publishing, 2016, pp. 490–496.
References [14] M. Han, Z. Li, J. S. He, D. Wu, Y. Xie, and A. Baba, “A novel
blockchain-based education records verification solution,” in
[1] S. Nakamoto. Bitcoin: A peer-to-peer electronic cash system. Proceedings of the 19th Annual SIG Conference on Information
Accessed April 4, 2022. [Online]. Available: http://bitcoin.org/ Technology Education, ser. SIGITE ’18. New York, NY, USA:
bitcoin.pdf Association for Computing Machinery, 2018, pp. 178–183.
[2] L. Hirtan, P. Krawiec, C. Dobre, and J. M. Batalla, “Blockchain- [Online]. Available: https://doi.org/10.1145/3241815.3241870
based approach for e-health data access management with [15] M. Turkanović, M. Hölbl, K. Košič, M. Heričko, and
privacy protection,” in 2019 IEEE 24th International Workshop A. Kamišalić, “Eductx: A blockchain-based higher education
on Computer Aided Modeling and Design of Communication credit platform,” IEEE Access, vol. 6, pp. 5112–5127, 2018.
Links and Networks (CAMAD). Limassol, Cyprus: IEEE, 11- [16] W. Gräther, S. Kolvenbach, R. Ruland, J. Schütte, C. Torres,
13 September 2019, pp. 1–7. and F. Wendland, “Blockchain for education: lifelong learning
[3] K. Christidis and M. Devetsikiotis, “Blockchains and smart passport,” in Proceedings of 1st ERCIM Blockchain workshop
contracts for the internet of things,” IEEE Access, vol. 4, pp. 2018, 2018: European Society for Socially Embedded Technolo-
2292–2303, 2016. gies (EUSSET). Amsterdam, Netherlands: European Society
[4] D. Salah, M. H. Ahmed, and K. ElDahshan, “Blockchain ap- for Socially Embedded Technologies (EUSSET), 2-8 May 2018,
plications in human resources management: Opportunities and pp. 1–8.
challenges,” in Proceedings of the Evaluation and Assessment [17] P. Ocheja, B. Flanagan, H. Ueda, and H. Ogata, “Managing
in Software Engineering. Trondheim, Norway: Association lifelong learning records through blockchain,” Research
for Computing Machinery, New York NY, United States, 15-17 and Practice in Technology Enhanced Learning, vol. 14,
April 2020, pp. 383–389. no. 1, p. 4, 2019. [Online]. Available: https://doi.org/10.1186/
[5] M. C. H. Clifton and S. Cox. (2018, January) ’staggering’ trade s41039-019-0097-0
in fake degrees revealed. Accessed April 4, 2022. BBC News. [18] R. A. Mishra, A. Kalla, A. Braeken, and M. Liyanage, “Privacy
[Online]. Available: https://www.bbc.com/news/uk-42579634 protected blockchain based architecture and implementation for
[6] L. M. Palma, M. A. Vigil, F. L. Pereira, and J. E. Martina, sharing of students’ credentials,” Information Processing and
“Blockchain and smart contracts for higher education registry in Management, vol. 58, no. 3, p. 102512, 2021.
brazil,” International Journal of Network Management, vol. 29, [19] M. Lodder and B. Zundel. Hyperledger indy
no. 3, p. e2061, May 2019. hipe. Accessed April 4, 2022. [Online]. Avail-

13
INTERNATIONAL JOURNAL OF INFORMATION SECURITY SCIENCE
O. Oksuz, Vol.11, No.1, pp.1-14

able: https://hyperledger-indy.readthedocs.io/projects/hipe/en/
latest/text/0109-anoncreds-protocol/README.html
[20] J. Camenisch and A. Lysyanskaya, “An efficient system for non-
transferable anonymous credentials with optional anonymity
revocation,” in Advances in Cryptology — EUROCRYPT 2001,
B. Pfitzmann, Ed. Berlin, Heidelberg: Springer Berlin Heidel-
berg, 2001, pp. 93–118.
[21] J. Camenisch and A. Lysyanskaya, “A signature scheme with
efficient protocols,” in Security in Communication Networks,
S. Cimato, G. Persiano, and C. Galdi, Eds. Berlin, Heidelberg:
Springer Berlin Heidelberg, 2003, pp. 268–289.
[22] R. L. Rivest, A. Shamir, and L. Adleman, “A method for
obtaining digital signatures and public-key cryptosystems,”
Communications of the ACM, vol. 21, no. 2, pp. 120–126, 1978.
[23] P. Paillier, “Public-key cryptosystems based on composite de-
gree residuosity classes,” in International conference on the
theory and applications of cryptographic techniques, J. Stern,
Ed. Prague, Czech Republic: Springer Berlin Heidelberg, 2-6
May 1999, pp. 223–238.
[24] C.-P. Schnorr, “Efficient signature generation by smart cards,”
Journal of cryptology, vol. 4, no. 3, pp. 161–174, 1991.
[25] Ethereum. https://www.ethereum.org. Accessed April 5, 2022.
[Online]. Available: https://www.ethereum.org
[26] M. Hanley and H. Tewari, “Managing lifetime healthcare
data on the blockchain,” in IEEE SmartWorld, Ubiquitous
Intell. Comput., Adv. Trusted Comput., Scalable Comput. Com-
mun., Cloud Big Data Comput., Internet People Smart City
Innov. (SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/SCI),
Guangzhou, China, 8-12 October 2018, pp. 246–251.

14