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

net/publication/316899123

Access control and data security in online document verification system

Conference Paper · December 2016

DOI: 10.1109/ICCIC.2016.7919707

CITATIONS READS
2 337

4 authors, including:

Ravinder Reddy B Rajrupa Singh

CVRS College of Engineering VIT University
4 PUBLICATIONS   2 CITATIONS    5 PUBLICATIONS   3 CITATIONS   

SEE PROFILE SEE PROFILE

Selvakumar Ramachandran
VIT University
48 PUBLICATIONS   79 CITATIONS   

SEE PROFILE

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

Reliable and secure space communication View project

IoT based Cloud Integration Security View project

All content following this page was uploaded by Ravinder Reddy B on 11 May 2018.

The user has requested enhancement of the downloaded file.

 Access Control and Data Security in Online
Document Verification System
Ravinder Reddy B∗ , Pavan Kumar C† , Rajrupa Singh‡ and Selvakumar R‡
∗ Department
of Computer Science and Engineering,
Anurag Group of Institutions, Hyderabad, Telangana
[email protected]
† School of Computer Science and Engineering (SCOPE), VIT University, Vellore

[email protected]
‡ School of Advanced Sciences (SAS), VIT University, Vellore

{rajrupa.singh,rselvakumar}@vit.ac.in

Abstract—Document or certificate verification is a crucial part available to access 24 × 7. But the challenge with such server
wherein a document or certificate issued by an authority will is to ensure the integrity and security of data from possible
be verified for its authenticity. In a typical system it will be attackers and hackers from changing and misusing the data
performed through exchange of mails or post, it is a time
consuming process and not foolproof due to human intervention. content hosted on these servers [1].
In an online system, the documents hosted on a third party server
with negligible downtime guarantee will be downloaded and A. Need for Access control in document verification system
verified. But in such cases the trustworthiness of the documents
hosted on these servers is questionable as the documents will Some of the online document verification system available
be hosted in its original format and there are chances that such as directverify owned by easydocs [2] will ask the user
it will be modified or altered. Trusting and protecting the
documents from getting modified due to intentional attacks to upload his/her document and the uploaded documents will
and other possible attacks are the issues to be addressed in be verified against University records by the above mentioned
online document verification system. To overcome these issues online document verification system providers, there by con-
and design a foolproof system, in this paper, we propose an stituting a semi online document verification system. Few
online document verification system based on Attribute Based online document verification system provides online document
Encryption (ABE). Third party server considered in this case
can be either a cloud or a server taken from a service provider. verification system wherein data owners i.e., the certificate
This paper is aimed at providing access control for users to access issuing authority will not have full control over the data hosted
the documents hosted online based on user’s attributes. We show on such systems as they are prone to insider attackers.
that the proposed scheme is robust against collusion attacks and In one hand web servers with reliable service uptime are
hacking kind of attacks. Also, it encourages the authorities such
available plenty in market to opt for with a reasonable cost.
as Universities, Autonomous institutions issuing documents for
adopting the proposed Online Document Verification System. On the other hand, cloud computing has emerged as a promis-
ing platform for providing storing and computation facilities
Keywords—Decentralized Attribute based encryption, Online
Document Verification System, Access control
on demand [3]. Cloud provides enormous computation and
storage capability for users with pay-as-you-use model and
I. I NTRODUCTION negligible downtime making it more attractive for both small
and large enterprises to opt for its services without any
Document or certificate verification has been a challenge hassle of maintaining physical servers on their premises. Data
wherein the certificate or document issued by an authority will stored on cloud or web servers will be stored in data centers
be checked or verified for its authenticity. The words certificate owned and maintained by cloud service providers at various
and document are used interchangeably and not to confuse geographical locations. As a result, users are not aware where
the word certificate with the cryptographic certificate. In exactly data is being stored and can’t exercise complete control
traditional document verification process, the authority issuing over their data. Another factor of concern for users on cloud
certificate will be contacted through email/mail for authenti- platform is the security and privacy of data as it may contain
cating the documents from the existing records of the issuing sensitive information. To ensure privacy and security concerns,
authority. It is time consuming process and also not fully users can opt for cryptographic principles and store encrypted
foolproof due to human intervention and other factors such as information on cloud making it accessible only for authorized
man-in-the-middle kind of attacks changing the contents of the users.
mail, whereas in the online document verification system, the In the recent era of information communication, most of
documents will be hosted on third party server (either cloud or the services offered by authorities, Government or Private,
web hosting) or server owned by the authorities. Such server is are availbale online and designing efficient and secure system
978-1-5090-0612-0/16/$31.00 ⃝
c 2016 IEEE to provide these services are the need of hour. Some of the
 2016 IEEE International Conference on Computational Intelligence and Computing Research

challenges in designing online document verification system TABLE I

are as follows: N OTATIONS

1) Requirement 1: Symbols Meanings

Ensuring security of the documents or certificates hosted AO Accounts Authority
CoE Data Owner
online so that it wont be available for unintended and PK Public Key of AO
unauthorized users. SK Secret Key of AO
2) Requirement 2: SK{i,GID} Secret key of User with GID
Cloud or web server should not change the contents of M Message
C Ciphertext
the documents hosted on them. H Hash function
3) Requirement 3: STUDID Student Identity Number
Users should not be given with write access to modify GID Global Identification Number
the content of documents and they can only have read
permission.
4) Requirement 4: B. Access Policies
Negligeble downtime should be guaranteed to users. Access policy to access data can be specified by boolean
To address these challenges, we propose Attribute Based function or a LSSS matrix by the data owner. Boolean function
Encryption (ABE) based online document verification system. can be equivalently represented using access tree with AN D
Leuko and Waters [4] proposed decentralized attribute based (∧) and OR (∨) as nodes and attributes are indicated as leaf
encryption scheme in which any entity can become an author- nodes [4]. Access policy specifies the permission to access the
ity by issuing a set of public and private keys with access poli- data based on the permission.
cies. Bethencourt et. al [5] proposed Ciphertext based attribute
encryption (CP-ABE) method wherein ciphertext is associated C. Multi-Authority Decentralized Attribute Based Encryption
with the set of access policies and keys are associated with the Scheme
users. CP-ABE scheme application is shown to be efficient in
Leuko and Waters [4] proposed Multi-Authority Decentral-
social networking to specify access permission [6], in cloud
ized Attribute based Encryption scheme in which any entity
computing paradigm [7] etc., to achieve user privacy and data
can become an authority by issuing set of public and private
security.
keys. There can be more than one entity which can become
We apply techniques from Leuko and Waters scheme to
authority and also there is no need of interdependency among
design attribute based Online document verification system.
the authorities (or KDCs) which was not the case in scheme
Also, we use techniques from Ruj et al. [8] aimed at providing
proposed by Chase and Chow [9].
distributed access control in cloud computing paradigm. But
Leuko and Waters [4] method consists of following algo-
the design in [8] is generic in nature and in present work, we
rithms:
design the system particularly in the context of constructing
online based document verification system where the docu- 1) Setup: A group G of prime order q generated by g,
ments are hosted either on cloud server or third party server. a bilinear map e : G × G → G are chosen. A hash
An additional condition we assume in our design is that both function H : {0, 1}∗ → G mapping the identities to G.
the entity Accounts and Cloud/third party server are using H is modeled as a random oracle.
different hosting provider (either cloud or web server). This An entity willing to become an authority will initialize
is to ensure that the entities and keys are not available for the the system by taking the security parameter λ and out-
hosting provider to access the information. puts global parameter. Entity takes the global parameter
The paper is structured as follows, in section II preliminaries as input and produces Secret key SK and Public key
required for establishing the proposed system is described. In P K. Further, entity will choose two random exponents
section III proposed system model and framework is discussed. αi , yi ∈ ZN and publishes its (entity’s) public key and
In section IV, document verification system workflow is given. secret key as follows:
In section V, analysis of proposed scheme is discussed. Section
VI deals with conclusion and future scope. P K = {e(g, g)αi , g yi ∀i} (1)

II. P RELIMINARIES
A. Bilinear pairing SK = {αi , yi ∀i} (2)
Let G be a multiplicative cyclic group of order q generated 2) Encrypt: Algorithm encrypt will take data D, access
by g. Map e is defined as e : G × G → G satisfying the matrix S, mapping attribute ρ and public key P K of
properties: relevant authority and produces ciphertext C.
1) Bilinearity: e(aP, bQ) = e(P, Q)ab for all P, Q ∈ G, a) A random s ∈ ZN and a random vector v ∈ ZN l
a, b ∈ Zq and Zq = {0, 1, 2, . . . , q − 1} with s as its first entry is choosen.
2) Non-degeneracy: Map is non degenerate: e(g, g) ̸= 1 b) Let λx = Sx · v, with Sx being row x of S.
 2016 IEEE International Conference on Computational Intelligence and Computing Research

c) Chose a random vector w ∈ ZN l

with 0 as its first Among the entities the Accounts office entity will act as
entry. l is the number of rows of matrix S and the an Authority by providing public and secret keys to users.
number of leaves in the access tree. The entity controller of Examination will be data owner and
d) Compute ωx = Sx · w the end users are the ones who want to access the documents
e) For rows of Sx of S, choose a random sx ∈ ZN hosted in ciphertext format on cloud/third party server entity.
Ciphertext is computed as follows ∀x: Cloud/third party server is the entity on which controller of
examination entity will host the encrypted document along
C0 = De(g, g)s
with access policy. The proposed model of online document
C1,x = e(g, g)λx e(g, g)αρ(x)rx verification system is given in Figure 1.
(3)
C2,x = g rx
A. Entities and their role
C3,x = g yρ(x) rx g ωx
1) Accounts Office (AO): AO entity plays a major role in
ρ is mapping of rows of access matrix S to its the proposed system. It takes the security parameter λ
attributes. and output global parameter N and g for the system. AO
f) Ciphertext sent by the sender along with access entity produces a pair of public key (P K) and secret key
matrix S and mapping of rows of access matrix to (SK) with global parameters.
its attributes (ρ) is given by, 2) Controller of Examination (CoE): Entity CoE being the
owner of data decides the access policy to access the
C =< S, ρ, C0 , {C1,x , C2,x , C3,x } > (4) data by users. It takes public key (P K) from accounts
3) KeyGen: Algorithm KeyGen generates corresponding and along with access policy and global parameters, it
secret keys to the users. Secret key is computed as produces a ciphertext C. Ciphertext C will be stored in
follows: third party/cloud server.
y 3) Cloud/third party server: This entity stores the cipher-
SKi,GID = g αi H(GID) i (5)
text C created by data owner i.e., CoE and makes it
where, GID is the global identifier of user and i is the available to users 24 × 7. Also, it ensures that the data is
attribute from the relevant authority. accessible only to users with suitable access permission.
4) Decrypt: Algorithm Decrypt takes ciphertext C, secret This entity is not allowed to modify or alter the data
key SKi,GID and GID of user as input and provide hosted on it.
message D as output. It is assumed that access matrix 4) Users: Entity users are the individuals or the government
(S, ρ) is provided to receiver along with the ciphertext. agencies who would like to access the documents hosted
The secret key possessed by the receiver should be on cloud/third party servers for verification. Users need
such that (1, 0, . . . , 0) is in the span of rows Sx of S, to pay a prescribed fee to the accounts and obtain a
otherwise decryption is not possible. token to be able to access the data.
For each row x is decrypted as follows:
B. Access Policy
C1,x · e(H(GID), C3,x ) ω Access policy is specified or defined by data owner CoE.
= e(g, g)λx e(H(GID), g) x
e(skπ(x),GID , C2,x ) Access policy for the proposed system is ”Any individual or
∑ (6) organization with Student ID and Token can access the data”.
Decryptor selects cx ∈ Zn such that x cx Sx = Access policy is written in the form of boolean attributes for
(1, 0, . . . , 0) and computes the proposed online document verification system as follows:
∏ ω cx s (TOKEN AND (STUDID AND (INDIVIDUAL OR GOVTA-
(e(g, g)λx e(H(u), g) x ) = e(g, g) (7)
GENCY)))
x
Any access tree can be converted into an equivalent LSSS
Since, λx = Sx · v, ωx = Sx · w, v · (1, 0, . . . , 0) = s and matrix. Procedure to convert access tree to an equivalent LSSS
w · (1, 0, . . . , 0) = 0. matrix is given in [4] Access tree in Figure 2 is given by
Data D is computed as equivalent matrix S as following:
C0  
D= s (8) 1 1 0
e(g, g) 0 −1 1 
S= 0 0 −1

III. S YSTEM M ODEL AND F RAMEWORK
0 0 −1
The following entities are involved in the proposed online
document verification system in an University setup. IV. P ROPOSED S YSTEM W ORKFLOW
1) Controller of Examination (CoE) Accounts entity AO will initialize the system and becomes
2) Accounts Office (AO) authority by taking global parameter as input and produces
3) Cloud/Third party server secret key and public key as follows: AO selects group
4) Users G of prime order q generated by q and a bilinear map
 2016 IEEE International Conference on Computational Intelligence and Computing Research

Fig. 1. Proposed Online Based Document Verification System

3) It chooses a random vector w ∈ Zql with 0 as its first

entry and compute ωx = Sx · w.
4) For each row Sx of S, choose a random sx ∈ ZN
5) Ciphertext of document D is calculated as follows:
C0 = De(g, g)s
C1,x = e(g, g)λx e(g, g)αρ(x)rx
(9)
C2,x = g rx
C3,x = g yρ(x) rx g ωx
U sers in need of document provided by CoE will approach
AO, obtain secret key to access the document and have
STUDID to access the document. Also, token by paying
necessary prescribed fee for accessing the document. AO
authority will provide secret key by taking global parameter
of system and GID of user to provide secret key for attribute
1 as follows:
Fig. 2. Access tree for the proposed online document verification system
SK1,GID = {g α1 H(GID)y1 } (10)

e : G × G → G. Bilinear Pairing over elliptic curve groups U sers can access data from the cloud/third party server.
is assumed in construction similar to Ruj et al. [8]. It chooses To decrypt the document, users should possess the SK
two random exponents αi and yi ∈ Zn and publishes P K = provided by AO and the keys possessed by the user for subset
{e(g, g)α1 , g y1 } and SK = {α1 , y1 } similar to equation 1 and of rows of access matrix S should be such that these rows are
2 respectively. spanned by (1, 0, · · · , 0), then the decryptor works as follows:
CoE will take the public key P K issued by AO and global For each such x, decryptor computes
parameters. CoE defines the access matrix generated from
Access tree in Figure 2. C1,x · e(H(GID), C3,x ) ω
= e(g, g)λx e(H(GID), g) x
CoE entity will blind the data D i.e., the document as e(skρ(x),GID , C2,x )
follows: ∑
Decryptor chooses constants cx ∈ Zn such that x cx Sx =
1) It chooses a random s ∈ Zq and a random vector v ∈ (1, 0, · · · , 0) and computes
ZlN , with s as its first entry. ∏ ω cx s
2) It calculates λx = Sx · v where A is the access matrix (e(g, g)λx e(H(u), g) x ) = e(g, g)
and Sx is row of matrix S. x
 2016 IEEE International Conference on Computational Intelligence and Computing Research

V. A NALYSIS OF P ROPOSED SYSTEM provide enormous bandwidth support and can easily mitigate
Proposed system satisfies the requirements of Decentralized the effect of DDoS attack. In the case of third party stand
Attribute based encryption scheme defined by Leuko and alone servers mitigating DDoS attack is a difficult task and
Waters. cannot effectively bypass such attacks.
Leuko and Waters [4] defined Decentralized Attribute based C. Collusion Attack
Encryption scheme as follows: ”A multi-authority CP-ABE
system is said to be correct if whenever GP is obtained from Users of the system will get the information what they are
the global setup algorithm, CT is obtained from the encryption entitled to and are not supposed to get additional information
algorithm on the message M, and {Ki,GID } is a set of keys even if more than one user colludes. Since each user is given
obtained from the key generation algorithm for the same with separate SK and e(H(GID), g) even if they collude they
ω
identity GID and for a set of attributes satisfying the access cannot efficiently compute e(H(GID), g) x .
structure of the ciphertext, Decrypt(CT, GP, {Ki,GID }) = VI. C ONCLUSION
M”. A secure and attribute based access system is proposed for
In the proposed model, GP is obtained from global set
achieving access control and data security in online document
up algorithm. Ciphertext C is obtained by entity CoE for
verification system with emphasis on Universtiy setup. The
message M by obtaining public key from authority AO. Key
proposed model meets the requirement of the present day
{SK1,GID } is the key obtained from AO authority by user
online document verification system and provides better access
with GID to decrypt data as given in equation 10. Thus,
control and full data ownership to the entities owning data.
the proposed system meets the requirement of online based
Also, it resists man-in-the-middle kind of attacks. As a future
document verification system.
scope, user revocation method and document authentication
A. Requirement fulfilment (that it is from relevant entity itself) can be added to make
Proposed scheme meets the requirement of online based the documents available to users in a dynamic group and on
document verification system outlined in section I as follows: temporal basis.
1) Requirement 1: Unauthorized Access R EFERENCES
Unauthorized access to data is not possible as to be [1] G. D. Gollin, “Verification of the integrity and legitimacy of academic
able to access the data, intruder or unauthorized user credential documents in an international setting,” College and University,
should possess set of rows which is a subset of access vol. 84, no. 4, p. 75, 2009.
[2] “Directverify.in,” 2016. [Online]. Available: https://www.directverify.in/
matrix and is a linear combination
∑ of (1, 0, · · · , 0) and myeasydocs directverify/directverify/index new.aspx
constant x ∈ Zq such that x cx R x = (1, 0, · · · , 0). [3] R. Buyya, C. S. Yeo, S. Venugopal, J. Broberg, and I. Brandic, “Cloud
From which, the message M , i.e., blinded by e(g, g)s computing and emerging it platforms: Vision, hype, and reality for
delivering computing as the 5th utility,” Future Generation computer
cannot be calculated. systems, vol. 25, no. 6, pp. 599–616, 2009.
2) Requirement 2: Secure Storage [4] A. Lewko and B. Waters, “Decentralizing attribute-based encryption,”
Message M is blinded and only ciphertext C is hosted in Annual International Conference on the Theory and Applications
of Cryptographic Techniques. Springer, 2011, pp. 568–588. [Online].
on cloud or third party server. Thus, cloud or third party Available: https://eprint.iacr.org/2010/351.pdf
server cannot know the content of documents. Server can [5] J. Bethencourt, A. Sahai, and B. Waters, “Ciphertext-policy attribute-
be curious and can only know the existence of document based encryption,” in 2007 IEEE symposium on security and privacy
(SP’07). IEEE, 2007, pp. 321–334.
but not the content of it. [6] S. Jahid, P. Mittal, and N. Borisov, “Easier: Encryption-based access
3) Requirement 3: No write access to users control in social networks with efficient revocation,” in Proceedings of
Proposed scheme provides read only access i.e., to the 6th ACM Symposium on Information, Computer and Communications
Security. ACM, 2011, pp. 411–415.
decrypt data based on user’s private key and users are [7] N. Balani and S. Ruj, “Temporal access control with user revocation
not allowed to write or do any modifications to the for cloud data,” in 2014 IEEE 13th International Conference on Trust,
documents. Security and Privacy in Computing and Communications. IEEE, 2014,
pp. 336–343.
4) Requirement 4: Negligible downtime [8] S. Ruj, A. Nayak, and I. Stojmenovic, “Dacc: Distributed access control
Negligible downtime is not a proposed model issue but in clouds,” in 2011IEEE 10th International Conference on Trust, Security
rather a suitable choice of reliable hosting provider or and Privacy in Computing and Communications. IEEE, 2011, pp. 91–98.
[9] M. Chase and S. S. Chow, “Improving privacy and security in multi-
cloud is essential to ensure this as many computations authority attribute-based encryption,” in Proceedings of the 16th ACM
are performed on cloud/third party server platform. conference on Computer and communications security. ACM, 2009, pp.
Apart from the requirement specified in section I, system is 121–130.
analyzed for the following attacks:
B. Distributed Denial of Service (DDoS)
Distributed Denial of Service attack is making the legitimate
user not to access the system by jamming the bandwidth
posing as actual user requesting for the resource. Cloud server
can mitigate such attacks very easily as cloud infrastructure

View publication stats